MINITAB Release 13 Session Command Quick Reference 1 MEET MTB UGUIDE 1 UGUIDE 2 CONTENTS HOW TO USE Session Command Quick Reference Using Session Commands You can use session commands for submitting commands interactively in MINITAB, or you can use session and macro commands for building macros (See Global and Local Macros (%macros)). Submitting session commands directly into MINITAB requires typing the commands one of two places: ■ The Session window ■ The Command Line Editor Session Window ■ Type commands at the MTB> prompt, subcommands at the SUBC> prompt, and data at the DATA> prompt. ■ By default, press e to execute a command and c+e to insert a blank line (these preferences can be changed). ■ If you are unable to type commands in the Session window, command language may be disabled. See Using Session Commands and Macros in MINITAB User’s Guide 1. Blue text denotes commands, subcommands, and arguments new to MINITAB Release 13.
217
Embed
MINITAB Release 13 Session Command Quick Reference · PDF fileMINITAB Release 13 Session Command Quick Reference 1 ... User’s Guide 1. Blue text denotes ... MINITAB [K] EXCEL PORTABLE
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
You can use session commands for submitting commands interactively in MINITAB, or you can use session and macro commands for building macros (See Global and Local Macros (%macros)). Submitting session commands directly into MINITAB requires typing the commands one of two places:
■ The Session window
■ The Command Line Editor
Session Window
■ Type commands at the MTB> prompt, subcommands at the SUBC> prompt, and data at the DATA> prompt.
■ By default, press e to execute a command and c+e to insert a blank line (these preferences can be changed).
■ If you are unable to type commands in the Session window, command language may be disabled. See Using Session Commands and Macros in MINITAB User’s Guide 1.
Blue text denotes commands, subcommands, and arguments new to MINITAB Release 13.
■ Open the Editor with c+L or Edit ➤ Command Line Editor.
■ Type commands at the cursor, then click Submit Commands to execute them.
■ You can also use, by default, c+e to execute a command and e to insert a blank line (these preferences can be changed).
■ Use the Command Line Editor to execute commands used previously in your session:
– Highlight commands in the Session window or History folder, then open the Command Line Editor. The highlighted commands appear in the editing window.
– Or, while the Session window or History folder is active, choose Edit ➤ Select All (c+A). When you open the Editor, all commands used in your current session will appear in the editing window. Since it is so easy to delete and edit commands in the editing window, this is a quick way to re-execute commands.
■ Press 3 to clear all commands.
■ You can cut, copy, and paste to and from the Command Line Editor, word processors, spreadsheets, and MINITAB windows—including the Data window.
NotationIn this Quick Reference, commands are left justified, subcommands are indented.K denotes a constant such as 8.3 or K14C denotes a column such as C12 or 'Height'E denotes either a constant or column, and sometimes a matrixM denotes a matrix such as M5[ ] denotes an optional argumentC…C means you can list one or more columns
Rules for Typing Commands■ You need only type the first four letters of any command. For %macros, type
% followed by the full macro name.
■ Start each command and subcommand on a new line.
■ To use a subcommand, end the line before it with a semi-colon (;). End the last subcommand with a period.For example: RANDOM 10 C1;
BINOMIAL 20 0.3.
■ Type the subcommand ABORT to cancel a multi-line session command.For example: RANDOM 10 C1;
■ Abbreviate lists of consecutive variables with a dash.For example: RANDOM 10 C1-C5
■ Enclose file names with double quotation marks. For example: SAVE "SALESQ1"
■ Enclose variable names with single quotation marks. For example: PLOT 'Sales'∗ 'Advertis'
■ Only type the commands and their arguments (K, C, E, or M as defined above). Text is for explanation only. For example, the syntax for the MEDIAN command is
MEDIAN of the values in C [put into K]This means that you could type either:
MEDIAN C1 to display the median of C1, orMEDIAN C1 K1 to display the median of C1 and store it in the constant K1
Interrupting CommandsTo stop the display of a command, press c+k.
OH = K number of lines for height of outputOW = K number of spaces for width of outputIW = K number of spaces for width of inputBRIEF = K controls amount of output from ARIMA, BBDESIGN,
CCDESIGN, CLUOBS, CLUVARS, DISCRIMINANT, FACTOR, FFDESIGN, GLM, KMEANS, MIXREG, REGRESS, RLINE, RSREG, SCDESIGN, and SLDESIGN. BRIEF 2 is the default; BRIEF 0 suppresses all output for all commands.
TSHARE operate interactivelyBATCH operate in batch mode
Files, Data, and Output CommandsNAME for E is 'name' … for E is 'name'
NEW open a new MINITAB worksheetWORKSHEET K (worksheet name)
CLOSE close a worksheetNOPROMPT do not prompt to save
CURRENT make the named worksheet the current worksheet
JPEG save in JPEG formatPNGB save in PNG black-and-white formatPNGC save in PNG color formatPNGH save in PNG high color formatTIFB save in TIF black-and-white formatTIFC save in TIF color formatBMPB save in BMP black-and-white formatBMPC save in BMP color formatBMPH save in BMP high color format
SAVE [in "filename"] a copy of the worksheetPORTABLE save a MINITAB portable worksheetREPLACE replace existing fileNOREPLACE do not replace existing fileSINGLE save in Release 10 (single precision) format
READ data into C…CFILE "filename"FORMAT (format statement)NOBS = KSKIP K linesDECIMAL "." or ","With READ; FILE only:TAB columns are separated with tabs
NONAMES columns are not namedALPHA columns are in K…K
DECIMAL "." or ","REPLACE replace existing fileNOREPLACE do not replace existing file
OUTFILE "filename" put all input and output in fileOW = K output width of fileOH = K output height of fileNOTERM no output to terminalNOECHO display no commands in outfile
NOOUTFILE output to terminal only; ends OUTFILENEWPAGE start next output on a new pageJOURNAL ["filename"] record MINITAB commands in this fileNOJOURNAL cancels JOURNALMTITLE K or "title" (display custom title above output in Session window)
Editing and Manipulating DataLET C(K) = K (changes the number in row K of C)DELETE rows K…K of C…C or C…CERASE all data in E…EINSERT data [between rows K and K] of C…C
FILE "filename"FORMAT (format statement)NOBS = KSKIP K linesDECIMAL "." or ","
COPY C…C into C…CCOPY C into K…K
USE rows K…KUSE rows where C = K…KOMIT rows K…KOMIT rows where C = K…K
MERGE creates a new worksheet using columns of data from two worksheetsNAME new worksheet K or "text string"WORKSHEET to be merged is K or "worksheet name" (must be used twice)
BY use C…C to match rows in worksheetNOMULTIPLESNOUNMATCHEDINCLUDE C…C
MISSINGSVORDER C…C
VALUES K…K or CWORKSHEETALPHABETICAL
SUBSET creates a new worksheet using specified rows of the current worksheetNAME KINCLUDEEXCLUDEBRUSHEDWHERE expressionROWS K…KCOLUMNS C…CMATRICES M…MNOMATRICESNOCONSTANTS
SPLIT splits a worksheet into multiple new worksheets,
Input Data Type: Text – no subcommands available (alternatively, you can use ATON, below)
ATON input text column in C, output numeric column in C
NTOA input numeric column in C, output text column in CWIDTH K charactersMAXWIDTH K charactersSIGNIFICANT K digitsDECIMALS K decimal placesMISSING (value converted to) "text"
STACK (E…E)… on (E…E), put in (C…C)SUBSCRIPTS into C
USENAMES create subscripts based on variable namesNEW put results in new worksheet
USENAMES create subscripts based on variable namesNAME new worksheet K or "text string"
USENAMES create subscripts based on variable namesUNSTACK (C…C) into (E…E)… (E…E)
SUBSCRIPTS are in C…CMISSINGS include missing as a subscript value
VARNAMES name unstacked columns based on subscriptsNEW put results in new worksheetNAME new worksheet K or "text string"AFTER put results after last column in use
TRANSPOSE C to M or M to C or M to M (no new subcommands supported) TRANSPOSE C…C or M (new arguments and subcommands)
VARNAMES are in CSTORE in C…C or MLABELS store in CNEW put results in new worksheetNAME new worksheet K or "text string"AFTER put results after last column in use
%ROWTOC stack rows in C…C, put in CEXPAND values in C…CEXSTORE put in C…CSUBSCRIPTS put in C
CONVERT using table in C C, the data in C, and put in CCONCATENATE C…C put in CSORT C [carry along C…C] put into C [and C…C]
BY C…CDESCENDING C…C
RANK the values in C, put ranks into C
CENTER the data in C…C put into C…CLOCATION [subtracting K…K]SCALE [dividing by K…K]MINMAX [with K as minimum and K as maximum]
ArithmeticLET E = expressionExpressions may use
– Arithmetic operators + – ∗ / ∗∗ (exponentiation)– Comparison operators = ~= or <> > < <= >=– Logical operators & | ~and any of the following functions and column operations:
ABSOLUTE SQRT LOGTEN LOGE EXPONENTIATEANTILOG ROUND SIN COS TANASIN ACOS ATAN SIGNS NSCORESPARSUMS PARPRODUCTS COUNT N NMISSSUM MEAN STDEV MEDIAN MINIMUM
Examples: LET C2 = SQRT(C1 - MIN(C1))LET C3(5) = 4.5
Simple Arithmetic Operations
ADD E to E, …, to E, put into ESUBTRACT E from E put into EMULTIPLY E by E, …, by E put into EDIVIDE E by E put into ERAISE E to the power E put into E
Columnwise Functions
ABSOLUTE value of E put into ESQRT of E put into ELOGE of E put into ELOGTEN of E put into EEXPONENTIATE E put into E
MAXIMUM SSQ SORT RANK LAGEQ NE LT GT LEGE AND OR NOT CTIMENOW TODAY WHEN DATE TIMEPI (constant pi) E (constant E) MISS (constant ∗ ) CEILING FLOORGAMMA LNGAMMA IGAMMA DEGREES RADIANSRCOUNT RN RNMISS RSUM RMEANRSTDEV RMEDIAN RMINIMUM RMAXIMUM RRANGERSSQ
ANTILOG of E put into EROUND to integer E put into ESIN of E put into ECOS of E put into ETAN of E put into EASIN of E put into EACOS of E put into EATAN of E put into ESIGNS of E [put into E]PARSUMS of C put into CPARPRODUCTS of C put into C
Normal Scores
NSCORES of C put into C
Columnwise Statistics
COUNT the number of values in C [put into K]N (number of nonmissing values) in C [put into K]NMISS (number of missing values) in C [put into K]SUM of the values in C [put into K]MEAN of the values in C [put into K]STDEV of the values in C [put into K]MEDIAN of the values in C [put into K]MINIMUM of the values in C [put into K]
MAXIMUM of the values in C [put into K]RANGE of the values in C [put into K]SSQ (uncorrected sum of sq.) for C [put into K]
Rowwise StatisticsRCOUNT of E…E put into CRN of E…E put into CRNMISS of E…E put into CRSUM of E…E put into CRMEAN of E…E put into CRSTDEV of E…E put into CRMEDIAN of E…E put into CRMINIMUM of E…E put into CRMAXIMUM of E…E put into CRRANGE of E…E put into CRSSQ of E…E put into C
Indicator VariablesINDICATOR variables for subscripts in C, put into C…C
MatricesREAD [from "filename"] into a K by K matrix MPRINT M…M
TRANSPOSE C to M or M to C or M to M (no new subcommands supported) TRANSPOSE C…C or M (new arguments and subcommands)
VARNAMES are in CSTORE in C…C or MLABELS store in CNEW put results in new worksheetNAME new worksheet K or "text string"AFTER put results after last column in use
INVERT M into MDEFINE K into K by K matrix MDIAGONAL is C, put into MDIAGONAL of M, put into C
COPY C…C into MCOPY M into C…CCOPY M into M
USE rows K…KOMIT rows K…K
ERASE M…MEIGEN analysis for M put values into C [vectors into M]
In the following commands E can be either C, K, or M:
ADD E to E, put into ESUBTRACT E from E, put into E
Distributions and Random DataRANDOM K observations into C…C
BERNOULLI trials p = KT degrees of freedom = KF df numerator = K, df denominator = KCHISQUARE df = K
PDF for values in E…E [put results in E…E]T degrees of freedom = KF df numerator = K, df denominator = KCHISQUARE df = K
CDF for values in E…E [put results in E…E]F df numerator = K, df denominator = K [K noncentrality param.]T degrees of freedom = K [K noncentrality parameter]CHISQUARE df = K [K noncentrality parameter]
INVCDF for values in E…E [put results in E…E]F df numerator = K, df denominator = K [K noncentrality param.]T degrees of freedom = K [K noncentrality parameter]CHISQUARE df = K [K noncentrality parameter]
RANDOM, PDF, CDF, and INVCDF have the subcommands:
BINOMIAL no. of trials = K, probability of success = KPOISSON µ = KINTEGER discrete uniform on integers K to KDISCRETE values in C and probabilities in CNORMAL [µ = K [σ = K]]UNIFORM [continuous on the interval K to K]CAUCHY [location = K [scale = K]]LAPLACE [location = K [scale = K]]LOGISTIC [location = K [scale = K]]LOGNORMAL [µ = K [σ = K]]EXPONENTIAL [mean = K]GAMMA first shape parameter = K second shape parameter = KWEIBULL first shape parameter = K second shape parameter = KBETA first shape parameter = K second shape parameter = KHYPERGEOMETRIC population size=K, number of successes=K, sample size=K
SAMPLE K rows from C…C put into C…CREPLACE sample with replacement
BASE for random number generator = K
Dynamic Data Exchange (DDE)XDADD "application" "topic" "item"
AXIS K [K K] or C [K K]all of its subcommands (see page 56)
TICK K [K…K or C] or C [K…K or C]all of its subcommands (see page 56)
GRID K or Call of its subcommands (see page 56)
REFERENCE K K…K or K C or C K…K or C Call of its subcommands (see page 56)
MARGIN K K K KMINIMUM K KMAXIMUM K KSCFRAME include ticks and reference lines in scale calculationSCANNOTATION include markers, lines, and polygons in scale calculationTRANSPOSE axesFRAME (default)NOFRAME [K]
Regions Subcommands for PLOT:
FIGURE [K K K K]all of its subcommands (see page 60)
MAXIMUM K KSCFRAME include ticks and reference lines in scale calculationSCANNOTATION include markers, lines, and polygons in scale calculationTRANSPOSE axesFRAME (default)NOFRAME [K]
Regions Subcommands for TSPLOT:
FIGURE [K K K K]all of its subcommands (see page 60)
DATA [K K K K]all of its subcommands (see page 60)
LEGEND [K K K K]all of its subcommands (see page 60)
POLYGON K K … K K or C Call of its subcommands (see page 53)
MARKER K K … K K or C Call of its subcommands (see page 53)
Frame Customization Subcommands for BOXPLOT:
AXIS K [K K] or C [K K]all of its subcommands (see page 56)
TICK K [K…K or C] or C [K…K or C]all of its subcommands (see page 56)
GRID K or Call of its subcommands (see page 56)
REFERENCE K K…K or K C or C K…K or C Call of its subcommands (see page 56)
MARGIN K K K KMINIMUM K KMAXIMUM K KSCFRAME include ticks and reference lines in scale calculationSCANNOTATION include markers, lines, and polygons in scale calculationTRANSPOSE axesFRAME (default)NOFRAME [K]
MARKER K K … K K or C Call of its subcommands (see page 53)
Frame Customization Subcommands for CONTOURPLOT:
AXIS K [K K] or C [K K]all of its subcommands (see page 56)
TICK K [K…K or C] or C [K…K or C]all of its subcommands (see page 56)
GRID K or Call of its subcommands (see page 56)
REFERENCE K K…K or K C or C K…K or C Call of its subcommands (see page 56)
MARGIN K K K KMINIMUM K KMAXIMUM K KSCFRAME include ticks and reference lines in scale calculationSCANNOTATION include markers, lines, and polygons in scale calculationTRANSPOSE axesFRAME (default)NOFRAME [K]
Regions Subcommands for CONTOURPLOT:
FIGURE [K K K K]all of its subcommands (see page 60)
REFERENCE K K…K or K C or C K…K or C Call of its subcommands (see page 56)
MARGIN K K K KMINIMUM K KMAXIMUM K KSCFRAME include ticks and reference lines in scale calculationSCANNOTATION include markers, lines, and polygons in scale calculationTRANSPOSE axesFRAME (default)NOFRAME [K]
Regions Subcommands for CHART:
FIGURE [K K K K]all of its subcommands (see page 60)
DATA [K K K K]all of its subcommands (see page 60)
LEGEND [K K K K]all of its subcommands (see page 60)
AXIS K [K K] or C [K K]all of its subcommands (see page 56)
TICK K [K…K or C] or C [K…K or C]all of its subcommands (see page 56)
GRID K or Call of its subcommands (see page 56)
REFERENCE K K…K or K C or C K…K or C Call of its subcommands (see page 56)
MARGIN K K K KMINIMUM K KMAXIMUM K KSCFRAME include ticks and reference lines in scale calculationSCANNOTATION include markers, lines, and polygons in scale calculationTRANSPOSE axesFRAME (default)NOFRAME [K]
Regions Subcommands for HISTOGRAM:
FIGURE [K K K K]all of its subcommands (see page 60)
DATA [K K K K]all of its subcommands (see page 60)
TEXT K K "text" or C C C MARKER K K,…, K K or C CLINE K K,…, K K or C CPOLYGON K K,…, K K or C CASPECT K KGSAVE save graph in "filename"[other graph commands]
Core Graphs: Customizing the Frame—Subcommands Continued
FRAME (default)NOFRAME [K] MARGIN K K K KMINIMUM K KMAXIMUM K KSCFRAME include ticks and reference lines in scale calculationSCANNOTATION include markers, lines, and polygons in scale calculationTRANSPOSE axes
Consult the tables of how the SIDE and PLACEMENT subcommands interact when using AXIS [AXIS with SIDE and PLACEMENT on page 58], TICK [TICK with SIDE and PLACEMENT on page 58], and REFERENCE [REFERENCE with SIDE and PLACEMENT on page 59].
MKM plot positions determined by modified Kaplan-Meier methodKM plot positions determined by Kaplan-Meier methodHJ plot positions determined by Herd-Johnson methodALLPT plot all points when there are tied failure timesAVEPT plot average (median) of tied points when there are tied failure
timesMAXPT plot maximum of tied points when there are tied failure times
Distributions
WEIBULL distributionEXTVALUE (extreme value) distributionEXPONENTIAL distributionLOGISTIC distributionLLOGISTIC (loglogistic) distributionNORMAL distributionLGNEXP (lognormal base EXP) distributionLGNTEN (lognormal base TEN) distribution
MLE maximum likelihoodLSXY least squares (failure time (X) versus rank (Y))PTILES are in K…KCONFIDENCE level is KCI display confidence intervals on plotTABLE display percentiles in Session window
TITLE for graph is "text"GSAVE save graph in "file"
Character GraphsCharacter (or typewriter-style) graphs display in the Session window and do not use any of the graphics subcommands described previously. You can save character graphs by saving the Session window contents; to save, make the Session window active and choose File ➤ Save Window As.
In the Session window or Command Line Editor, type the command GSTD (standard graphics) to enable character graphs. To re-enable high-resolution (professional graphics) commands, type GPRO. DOTPLOT and STEM-AND-LEAF character graphs are available in GSTD and GPRO mode. When you produce graphs through the menus, MINITAB switches modes automatically.
GSTD enable character graphicsGPRO enable high-resolution graphics
HISTOGRAM C…CINCREMENT = KSTART at K [end at K]BY CSAME scales for all columns
STEM-AND-LEAF display of C…CTRIM outliersINCREMENT = KBY C
BOXPLOT for CINCREMENT = KSTART at K [end at K]BY CLINES = KNOTCH [K% confidence] sign c.i.LEVELS K…K
PLOT C vs. CSYMBOL = 'symbol'TITLE = 'text'FOOTNOTE = 'text'YLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XINCREMENT = KXSTART at K [end at K]
Note Single quotes must be used around text strings with these subcommands.
MPLOT C vs. C, and C vs. C, and… C vs. CTITLE = 'text'FOOTNOTE = 'text'YLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XINCREMENT = KXSTART at K [end at K]
LPLOT C vs. C using tags in CTITLE = 'text'FOOTNOTE = 'text'YLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XINCREMENT = KXSTART at K [end at K]
TPLOT y in C vs. x in C vs. z in CTITLE = 'text'FOOTNOTE = 'text'YLABEL = 'text'XLABEL = 'text'YINCREMENT = K
Note Single quotes must be used around text strings with these subcommands.
Note Single quotes must be used around text strings with these subcommands.
Note Single quotes must be used around text strings with these subcommands.
Global and Local Macros (%macros)Global and local macros use the default file extension MAC and are invoked using the symbol %. You can include a path statement after %.
To execute a global macro, type%filename (e.g. %mymacro or %c:\research\mymacro)
To execute a local macro, type%filename arguments;subcommandname arguments;…subcommandname arguments.
MCONSTANT list of variables that are constantsMCOLUMN list of variables that are columnsMMATRIX list of variables that are matricesMFREE list of variables, type determined when macro executedDEFAULT argument = value … argument = value
Variables in Local Macros
KKCAT K K K (combine the text in the first two K’s, store combined text in the third K; all K’s are text constants)
KKNAME K C, K C, … (store the text constant K in column C)KKSET K "text", K "text", … (store the text in the text constant K)MTYPE variablename K (local macros only: if variable is a constant
returns K = 1; if a column K = 2; if a matrix K = 3)DTYPE E K returns the data type of a column or constant
LET function version:LET K1 = DTYPE(K2) stores the type for K2 in K1
Control Statements for Global and Local Macros
IF logical expression (same as for LET)ELSEIF logical expressionELSEENDIFDO K = list of numbers
ENDDOWHILE logical expressionENDWHILENEXT transfers control to beginning of DO or WHILE loopBREAK goes to the command that follows a DO or WHILE loopGOTO numberMLABEL numberCALL template (calls a subroutine in the same file)RETURN ends subroutine and returns control to main macroEXIT stop macro, return to MINITAB sessionPAUSE transfers control from macro to keyboardRESUME returns control to the macro
Using DOS Commands
CD [path] change directoryDIR [path] list file namesTYPE "[path]filename.ext" displays a file
Labeling Output of Global and Local Macros
TITLE (default; displays output titles in Session window)NOTITLE suppress output titles in Session windowWTITLE "title" for Graph window
ECHO displays macro commands in Session window during executionNOECHO (default)DEBUG tells how the macro is processingNODEBUG (default)TYPE "[path]filename.ext" displays a filePAUSE transfers control from macro to keyboardRESUME transfers control from keyboard back to macro
Handling Errors in Macros
PLUG macro keeps going when it finds an errorNOPLUG (default)MRESET restore environment settings after macro runsINFO displays summary of local worksheetSWAP fit macro into available memoryNOSWAP (default)
Additional Local Macro Features
WRITE ["filename"] C…C stores indicated local worksheet columns in atext file
BRIEF = K controls amount of output from ARIMA, BBDESIGN, CCDESIGN, CLUOBS, CLUVARS, DISCRIMINANT, EVDESIGN, FACTOR, FFACTORIAL, FFDESIGN, GLM, KMEANS, LREGRESSION, LTABLE, LTEST, MIXREG, OPTDES, PROBIT, REGRESS, RLINE, RSREG, SCDESIGN, and SLDESIGN
OH K (default = 0; controls pausing of output in Session window)GPAUSE [K]GPRINT with no arguments prints the newest Graph windowGSCALE K = minimum data value K = maximum data value
NMINIMUM K = minimum number of tick to useNMAXIMUM K = maximum number of ticks to useNTICKS store number of ticks in KTMINIMUM store minimum tick value in KTMAXIMUM store maximum tick value in KTINCREMENT store distance between ticks in KSMINIMUM store minimum scale value in KSMAXIMUM store maximum scale value in K
any graphics commandNOBRUSH disable brushing on the resulting graph
NOTE [comment] (displays blank lines or messages)YESNO K (returns K = 1 if user responds "Yes", K = 0 for "No")PAUSE transfers control from macro to keyboardRESUME transfers control from keyboard back to macro
ExecsExecs use the default file extension MTB and are invoked using the command EXECUTE.
EXECUTE ["filename"] [K times] (default filename = MINITAB.MTB)ECHO display the commands that followNOECHO do not display the commands that follow (default)YESNO store response ("Yes" = 1, "No" = 0) in K
Looping Through Columns and Matrices
Use the CK syntax to use a constant to specify a column number. If K1 = 3, then CK1 = C3; if K2 = 16, then CK2 = C16. Specify a matrix using the MK syntax. Here is an example of an Exec called PLOTS.MTB that uses CK:
Start the Exec and generate 20 plots by typing:LET K1 = 1EXECUTE "PLOTS" 20
Canceling a %Macro or ExecTo cancel Press c+k.
Basic StatisticsDESCRIBE variables in C…C
BY levels in CGHISTOGRAM display histogram for each variableGNHISTOGRAM display histogram with a normal curve for each variableGDOTPLOT display dotplot for each variableGBOXPLOT display boxplot for each variable
%DESCRIBE variables in C…CBY levels in CCONFIDENCE level is KTITLE for graph is "text"GSAVE save graph in "filename"
The first time through the loop, K1 = 1, giving PLOT C21 ∗ C1. The next time through the loop, K1 = 2, giving PLOT C21 ∗ C2, and so on.
ONEZ C…C one-sample Z-test and confidence intervalSIGMA KTEST KALTERNATIVE = KCONFIDENCE KGHISTOGRAM display histogram for each variableGDOTPLOT display dotplot for each variableGBOXPLOT display boxplot for each variable
ONET C…C one-sample t-test and confidence intervalTEST KALTERNATIVE = KCONFIDENCE KGHISTOGRAM display a histogramGDOTPLOT display a dotplotGBOXPLOT display a boxplot
TWOSAMPLE test and c.i. [K% confidence] samples in C CALTERNATIVE = KPOOLED procedureGDOTPLOT display a dotplotGBOXPLOT display a boxplot
TWOT test and c.i. [K% confidence] data in C, groups in CALTERNATIVE = KPOOLED procedure
STDEVS store in CVARIANCES store in CUPPER store in CLOWER store in CNOTABLE suppress table of standard deviations and confidence limitsTITLE for graph is K or "text"GSAVE save graph in K or "filename"
CORRELATION between C…C [put into M]NOPVALUES suppresses display of p-values
COVARIANCE between C…C [put into M]%NORMPLOT C
PROBS CSWTEST Ryan-Joiner goodness-of-fit test (sim. to Shapiro-Wilk)KSTEST Kolmogorov-Smirnov goodness-of-fit testTITLE for graph is "text"GSAVE save graph in "filename"
RegressionREGRESS C on K predictors C…C
CONSTANT fit y-intercept (default)NOCONSTANT do not fit y-interceptWEIGHTS are in C
SRESIDUALS put into CFITS put into CMSE put into KCOEFFICIENTS put into CXPXINV put into MRMATRIX put into MHI put into C (leverage)RESIDUALS put into C (observed - fit)TRESIDUALS put into C (deleted Studentized)COOKD put into C (Cook’s distance)DFITS put into CPREDICT for E…E
CONFIDENCE level is KPFITS put fits into CPSDFITS put standard deviation of fits into CCLIMITS put confidence limits into C CPLIMITS put prediction limits into C C
VIF display variance inflation factorDW display Durbin-Watson statisticPURE perform pure error lack-of-fit testXLOF perform experimental lack-of-fit testRTYPE plot residual type KGHISTOGRAM display histogram of residualsGNORMAL display normal plot of residuals
GFITS plot residuals versus fitsGORDER plot residuals versus orderGVARIABLE plot residuals versus C…CTOLERANCE K [K]BRIEF output code K
STEPWISE regress C on predictors C…CCONSTANT fit y-intercept (default)NOCONSTANT do not fit y-interceptFENTER = K (default is 4)FREMOVE = K (default is 4)FORCE C…CENTER C…CREMOVE C…CBEST K alternative predictorsSTEPS = K (default depends on output width)FORWARD sets F-to-remove to 0 and alpha-to-remove to 1BACKWARD F-to-enter to 10**18 and alpha-to-enter to 0AENTER K alpha-to-enterAREMOVE K alpha-to-remove
BREG C on predictors C…CCONSTANT fit y-intercept (default)NOCONSTANT do not fit y-interceptINCLUDE predictors C…C in all models
BEST K modelsNVARS K [up to K]NOWARN suppress warning for model with many predictors
%FITLINE regress y in C on the predictor in CPOLY order of polynomial is KLOGY use log transformation on responseLOGX use log transformation on predictorLOGSCALE display log scale for transformed variablesFITS put fits into CRESIDUALS put residuals into CLFITS put log-transformed fits into CLRES put residuals of log-transformed response into CCOEFFICIENTS put regression coefficients into CCI display confidence bandsPI display prediction bandsCONFIDENCE level is KTITLE for graph is "text"GSAVE save graph in "filename"
CONSTANT fit y-intercept in REGRESS, STEPWISE, and BREG commands thatfollow
NOCONSTANT do not fit y-intercept in REGRESS, STEPWISE, and BREG commands that follow
BLOGISTIC modelFREQUENCY are in CST response is number of successes and trialsSF response is number of successes and failuresFT response is number of failures and trialsLOGIT link functionNORMIT link functionGOMPIT link functionFACTORS are in C…CREFERENCE level or event for C is K, …, C KGPDCHISQUARE plot delta chi-square vs event probabilityGPDDEVIANCEplot delta deviance vs event probabilityGPDSBETA plot delta beta (standardized) vs event probabilityGPDBETA plot delta beta vs event probabilityGHDCHISQUARE plot delta chi-square vs leverageGHDDEVIANCE plot delta deviance vs leverageGHDSBETA plot delta beta (standardized) vs leverageGHDBETA plot delta beta vs leveragePRESIDUALS put in CSPRESIDUALS put in CDRESIDUALS put in CDCHISQUARE put in CDDEVIANCE put in CDSBETA put in C
DBETA put in CHI put in CEPROBABILITY put in C [C]NOCCUR put in C [C]NTRIALS put in CCOEFFICIENTS put in CSECOEFFICIENTS put in CXPWXINVERSE put in MLOGLIKELIHOOD put in KHOSMER number of groups is KBRIEF output code KSTEP display log-likelihood for each iterationITERATION maximum number is KSTART values in CCTOLERANCE convergence criteria is K KTOLERANCE level is K [K]
OLOGISTIC modelFREQUENCY counts are in CLOGIT link functionNORMIT link functionGOMPIT link functionFACTORS are in C…CREFERENCE level or event for C is K, …, C K
ORDER of response levels is C or K…KEPROBABILITY put in C…CCUMPROBABILITY put in C…CNOCCUR put in C…CNTRIALS put in CCOEFFICIENTS put in CSECOEFFICIENTS put in CXPWXINVERSE put in MLOGLIKELIHOOD put in KBRIEF output code KSTEP display log-likelihood for each iterationITERATION maximum number is KSTART values in CCTOLERANCE convergence criteria is K KTOLERANCE level is K [K]
NLOGISTIC modelFREQUENCY counts are in CFACTORS are in C…CREFERENCE level or event for C is K, …, C KEPROBABILITY put in C…CNOCCUR put in C…CNTRIALS put in CCOEFFICIENTS put in C
SECOEFFICIENTS put in CXPWXINVERSE put in MLOGLIKELIHOOD put in KBRIEF output code KSTEP display log-likelihood for each iterationITERATION maximum number is KSTART values in CCTOLERANCE convergence criteria is K KTOLERANCE level is K [K]
%RESPLOTS residuals in C, fits in CTITLE for graph is "text"GSAVE save graph in "filename"
Analysis of VarianceAOVONEWAY aov on samples in C…C
GDOTPLOT display a dotplotGBOXPLOT display a boxplot
ONEWAY aov on data in C, levels in C [put residuals in C [fits in C]]GDOTPLOT display a dotplotGBOXPLOT display a boxplotGHISTOGRAM display a histogram of the residualsGNORMAL display normal probability plot of the residuals
GFITS plot residuals vs fitted valuesGORDER plot residuals vs order of the dataGVARIABLE plot residuals vs C…CTUKEY [family error rate K]FISHER [individual error rate K]DUNNETT [family error rate K] for control group KMCB [family error rate K] best is K
TWOWAY aov on data in C, levels in C C [put residuals in C [fits in C]]ADDITIVE modelMEANS for the factors C [C]GHISTOGRAM display histogram of residualsGNORMAL display normal probability plot of residualsGFITS display plot of residuals vs fitsGORDER display plot of residuals vs orderGVARIABLE display plot of residuals vs C…C
%ANOM analysis of means on data in C, levels in C [C]ALPHA level is KTABLE display table of statisticsTITLE for graph is "text"GSAVE save graph in "filename"
%BANOM analysis of means on binomial data in C, sample size KALPHA level is KTABLE display table of statisticsTITLE for graph is "text"
GSAVE save graph in "filename"%PANOM analysis of means on Poisson data in C
ALPHA level is KTITLE for graph is "text"GSAVE save graph in "filename"
ANOVA modelRANDOM C…CEMS print table of expected mean squares, estimated variance
components, and error term in each F-testFITS put into C…CRESIDUALS put into C…CMEANS for termlistTEST for termlist / errortermRESTRICT use restricted modelGHISTOGRAM display histogram of residualsGNORMAL display normal probability plot of residualsGFITS display plot of residuals vs fitted valuesGORDER display plot of residuals vs order of dataGVARIABLE display plot of residuals vs C…CMANOVA [termlist [/ errorterm]]SSCP print hypothesis matrix and error matrixEIGEN print table of eigenvalues and eigenvectorsPARTIAL print matrix of partial correlations
ANCOVA modelCOVARIATES are in C…CFITS put into C…CRESIDUALS put into C…CMEANS for termlistTEST for termlist / errortermGHISTOGRAM display histogram of residualsGNORMAL display normal probability plot of residualsGFITS display plot of residuals vs fitted valuesGORDER display plot of residuals vs order of dataGVARIABLE display plot of residuals vs C…C
GLM modelCOVARIATES are in C…CWEIGHTS are in CFITS put into C…CRESIDUALS put into C…CSRESIDUALS put into C…CTRESIDUALS put into C…CHI put into CCOOKD put into C…CDFITS put into C…CXMATRIX put into M
COEFFICIENTS put into C…CMEANS for termlistTEST for termlist / errortermMANOVA [termlist [/ errorterm]]RANDOM C…CSSQUARES KEMSINTERACT C…CSMEANS store means in C…CPAIRWISE terms
TUKEYBONFERRONISIDAKCONFIDENCE KNOTESTNOCI
CONTROL termsLEVELS K K …KALTERNATIVE KDUNNETTBONFERRONISIDAKCONFIDENCE KNOTEST
NOCISSCP print hypothesis matrix and error matrixEIGEN print table of eigenvalues and eigenvectorsPARTIAL print matrix of partial correlationsNOUNIVARIATE suppress univariate outputRTYPE plot residual type KGHISTOGRAM display histogram of residualsGNORMAL display normal probability plot of residualsGFITS display plot of residuals vs fitted valuesGORDER display plot of residuals vs order of dataGVARIABLE display plot of residuals vs C…CBRIEF output code KTOLERANCE K [K]
%VARTEST response in C factors in C…CUNSTACKEDCONFIDENCE level is KSTDEVS store in CVARIANCES store in CUPPER store in CLOWER store in CNOTABLE suppress table of standard deviations and confidence limitsTITLE for graph is K or "text"GSAVE save graph in K or "filename"
YMIN y-axis minimum is KYMAX y-axis maximum is KTITLE for graph is "text"GSAVE save graph in "filename"
%INTERACT interaction plots for factors in C…CRESPONSE data in CLSMEANS data in CFULL display full matrix of plotsNOCENTER remove center points from factor level columnsYMIN y-axis minimum is KYMAX y-axis maximum is KTITLE for graph is "text"GSAVE save graph in "filename"
% RESPLOTS see %RESPLOTS under Regression, page 101
Multivariate AnalysisDISCRIMINANT groups in C, predictors in C…C
QUADRATIC use quadratic discriminant analysisPRIORS are in K…KLDF coefficients put in C…CFITS put in C [XVAL fits in C]XVAL use cross-validation
PCA principal component analysis of C…CCOVARIANCE matrixNCOMP number of components is KCOEFFICIENTS put into C…CSCORES put into C…CGSCREE display a scree plotGSCORE plot scores for first two components
FACTOR [C…C]CORRELATION factor the correlation matrix [M]COVARIANCE factor the covariance matrix [M]NFACTORS extract K factorsML use maximum likelihood (default is PCA) PRIORS use C as initial communality estimates in MLMAXITERATIONS maximum number of iterations for ML is KCONVERGENCE criterion for convergence in ML is KINLOADINGS input C…C as loadingsSORT sort loadings, [print all loadings less than K as 0.0]VMAX use varimax rotationQMAX use quartimax rotationEMAX use equamax rotationOMAX use orthomax rotation with γ = K
EIGEN store eigenvalues in C, eigenvectors in MLOADINGS store loadings in C…CCOEFFICIENTS store factor score coefficients in C…CSCORES store factor scores in C…CRESIDUAL store residual matrix in MROTMAT store rotation matrix in MGSCREE display a scree plotGSCORE plot scores for first two factorsGLOADING plot loadings for first two factorsBRIEF output code K
CLUOBS cluster observations in C…C or MEUCLIDEAN use Euclidean distance (default)SQEUCLIDEAN use squared Euclidean distancePEARSON use Pearson distanceSQPEARSON use squared Pearson distanceMANHATTAN use Manhattan distanceSINGLE use single linkage (default)AVERAGE use average linkageCOMPLETE use complete linkageMCQUITTY use McQuitty linkageCENTROID use centroid linkageMEDIAN use median linkageWARD use Ward linkage
STANDARDIZE variables before clusteringNUMBER cut dendrogram to obtain K clustersCUT cut dendrogram at K similarity levelDENDROGRAM display dendrogram
TYPE line type for each cluster is K…KCOLOR line color for each cluster is K…KSIZE line size for each cluster is K…KTITLE for dendrogram is "text"GSAVE save dendrogram in "filename"YDISTANCE use distance on y-axis, default is similarityMULTIPAGE display each cluster in a separate window
DMATRIX save the distance matrix in MMEMBERSHIP save cluster membership in CDISTANCE save distance between observations and centroids in C…CBRIEF output code K
CLUVARS cluster variables in C…C or MCORRELATION use correlation to measure distance (default)ABSCORRELATION use absolute correlation to measure distanceSINGLE use single linkage (default)AVERAGE use average linkageCOMPLETE use complete linkageMCQUITTY use McQuitty linkageCENTROID use centroid linkage
MEDIAN use median linkageWARD use Ward linkageNUMBER cut dendrogram to obtain K clustersCUT cut dendrogram at K similarity levelDENDROGRAM display dendrogram
TYPE line type for each cluster is K…KCOLOR line color for each cluster is K…KSIZE line size for each cluster is K…KTITLE for dendrogram is "text"GSAVE save dendrogram in "filename"YDISTANCE use distance on y-axis, default is similarityMULTIPAGE display each cluster in a separate window
DMATRIX save the distance matrix in MBRIEF output code K
KMEANS cluster observations in C…CSTANDARDIZE variables before clusteringNUMBER extract K clustersINITIAL use the initial partition in CMEMBERSHIP save cluster membership in CDISTANCE save distances between observations and centroids in C…CBRIEF output code K
COUNTS display frequenciesPERCENTS display percentage of total nonmissing valuesCUMCOUNTS display cumulative counts
CUMPERCENTS display cumulative percents ALL display all four statistics above
STORE in C…C
CHISQUARE test on table stored in C…C
TABLE the data classified by C…CMEANS for C…CMEDIANS for C…CSUMS for C…CMINIMUMS for C…CMAXIMUMS for C…CSTDEV for C…CSTATS for C…CDATA for C…CN for C…CNMISS for C…CPROPORTION of cases = K [through K] for C…CCOUNTS include total obs. count in each cell
ROWPERCENTS include row percent in each cellCOLPERCENTS include column percent in each cellTOTPERCENTS include total percents in each cellCHISQUARE analysis [output code = K]MISSING level for classification variable C…CNOALL in marginsALL for C…CFREQUENCIES are in CLAYOUT K rows by K columns
CA simple correspondence analysis of variables in C…CCONTINGENCY input consists of columns of a contingency tableRNAMES are in CCNAMES are in CNCOMPONENTS are KRCROSS cross the first two variables to form the rowsCCROSS cross the last two variables to form the columnsRSUPPLEMENTARY are in C…CCSUPPLEMENTARY are in C…CSRNAMES are in CSCNAMES are in CCTABLE print the contingency tableRPROFILES print the row profilesCPROFILES print the column profiles
EXPECTEDS print table of expected cell frequenciesDIFFERENCES print table of (observed-expected) in each cellCHISQUARE print table of chi-square value in each cellINERTIAS print table of relative inertia in each cellRPLOT display a row plotCPLOT display a column plotSPLOT display a symmetric plot of the rows and columnsARPLOT display an asymmetric row plotACPLOT display an asymmetric column plotAXES for plots are K K, …, K KSUPPLEMENTARY display supplementary points in each plotRPCOORDINATES put row principal coordinates into C…CRSCOORDINATES put row standardized coordinates into C…CCPCOORDINATES put column principal coordinates into C…CCSCOORDINATES put column standardized coordinates into C…CSVTB put the contingency table into C…C
MCA multiple correspondence analysis of variables in C…CINDICATORS columns on MCA are indicator variablesCNAMES are in CCSUPPLEMENTARY are in C…CSCNAMES are in CNCOMPONENTS are KITABLE print table of indicator variablesBURT print Burt table
AXES for plots are K K … K KSUPPLEMENTARY display supplementary points in all plotsCPLOT display column plotCOORDINATES put into C…C
Time Series%TREND analysis for the data in C
QUADRATIC use quadratic trend modelGROWTH use exponential growth trend modelSCURVE use Pearl-Reed logistic trend modelPRIORS use coefficients K K [K]WEIGHTS to compute smoothed coefficients are K K [K]FORECASTS generate K forecastsORIGIN begin forecasts from time KRESIDUALS store the residuals in CFITS store the fits in CFSTORE store the forecasts in CNOPLOT suppress the plotTABLE display additional outputTITLE for graph is "text"GSAVE save graph in "filename"
%DECOMP analysis for the data in C with seasonal length K
MULTIPLICATIVE model (default)ADDITIVE modelNOTREND do not include a trend componentSTART initial seasonal period is KFORECASTS generate K forecastsORIGIN begin forecasts from time KTREND store the trend component values in CDETREND store the detrended data in CSEASONAL store the seasonal component values in CDESEASONAL store the deseasonalized data in CFSTORE store the forecasts in CRESIDUALS store the residuals in CFITS store the fits in CNOPLOT suppress all plotsTABLE display additional outputTITLE for graph is "text"GSAVE save graph in "filename"
%MA for data in C, with a moving average of length KCENTER the moving averagesFORECASTS generate K forecastsORIGIN generate forecasts from time KAVERAGES store the moving averages in CFITS store the fits in CRESIDUALS store the residuals in C
FSTORE store the forecasts in CUPPER store the upper prediction limits in CLOWER store the lower prediction limits in CSMPLOT plot smoothed values vs actual dataNOPLOT suppress the plotTABLE display additional outputTITLE for graph is "text"GSAVE save graph in "filename"
%SES single exponential smoothing for the data in CWEIGHT to use to smooth the data is KINITIAL smooth value is average of first K valuesFORECASTS generate K forecastsORIGIN generate forecasts from time KSMOOTHED store the smoothed data in CFITS store the fits in CRESIDUALS store the residuals in CFSTORE store the forecasts in CUPPER store the upper prediction limits in CLOWER store the lower prediction limits in CSMPLOT plot smoothed values vs actual dataNOPLOT suppress the plotTABLE display additional outputTITLE for graph is "text"GSAVE save graph in "filename"
%DES double exponential smoothing for the data in CWEIGHTS for level component are K, for trend component are KFORECASTS generate K forecastsORIGIN generate forecasts from time KSMOOTHED store the smoothed data in CLEVEL store the level component values in CTREND store the trend component values in CFITS store the fits in CRESIDUALS store the residuals in CFSTORE store the forecasts in CUPPER store the upper prediction limits in CLOWER store the lower prediction limits in CSMPLOT plot smoothed values vs actual dataNOPLOT suppress the plotTABLE display additional outputTITLE for graph is "text"GSAVE save graph in "filename"
%WINTMULT multiplicative seasonal expo. smoothing for data in C withseasonal length K
WEIGHTS for level component = K, for trend component = K, for seasonalcomponent = K
FORECASTS generate K forecastsORIGIN generate forecasts from time KSMOOTHED store the smoothed data in C
LEVEL store the level component values in CTREND store the trend component values in CSEASONAL store the seasonal component values in CRESIDUALS store the residuals in CFITS store the fits in CFSTORE store the forecasts in CUPPER store the upper prediction limits in CLOWER store the lower prediction limits in CSMPLOT plot smoothed values vs. actual dataNOPLOT suppress the plotTABLE display additional outputTITLE for graph is "text"GSAVE save graph in "filename"
%WINTADD additive seasonal expo. smoothing for the data in C with seasonallength K
WEIGHTS for level component = K, for trend component =K,for seasonal component = K
FORECASTS generate K forecastsORIGIN generate forecasts from time KSMOOTHED store the smoothed data in CLEVEL store the level component values in CTREND store the trend component values in CSEASONAL store the seasonal component values in CRESIDUALS store the residuals in C
FITS store the fits in CFSTORE store the forecasts in CUPPER store the upper prediction limits in CLOWER store the lower prediction limits in CSMPLOT plot smoothed values vs. actual dataNOPLOT suppress the plotTABLE display additional outputTITLE for graph is "text"GSAVE save graph in "filename"
DIFFERENCES [of lag K] for data in C, put into CLAG [by K] for data in C, put into C
ACF [up to K lags] for series in C [put into C]PACF [up to K lags] for series in C [put into C]%ACF for series in C
MAXLAG maximum number of lags is KCORR store autocorrelations in CTSTATS store t-statistics in CLBQ store Ljung-Box Q statistics in CTITLE for graph is "text"GSAVE save graph in "filename"
%PACF for series in CMAXLAG maximum number of lags is KCORR store partial autocorrelations in C
TSTATS store t-statistics in CTITLE for graph is "text"GSAVE save graph in "filename"
CCF [up to K lags] between series in C and C
ARIMA p = K, d = K, q = K [P = K, D = K, Q = K, S = K] data in C[put residuals in C [predicted values in C [parameters in C]]]
CONSTANT fit y-intercept (default)NOCONSTANT do not fit y-interceptSTARTING values are in CFORECAST [origin = K] up to K leads [forecasts in C [limits in C C]]GSERIES display time series plotGACF plot autocorrelation function for residualsGPACF plot partial autocorrelation function for residualsGHISTOGRAM display histogram of residualsGNORMAL display normal probability plot of residualsGFITS plot residuals vs fitted valuesGORDER plot residuals vs order of dataGVARIABLE plot residuals vs C…CBRIEF output code K
Exploratory Data AnalysisLVALS of C [put letter values in C [mids in C [spreads in C]]]
MPOLISH C, rows in C, columns in C [put residuals in C [fits in C]]COLUMNS start iteration with column mediansITERATIONS = KEFFECTS put commons into K, rows into C, columns into CCOMPARISONS values, put into C
RLINE y in C, x in C [put residuals in C [fits in C [coefficients in C]]]MAXITERATIONS maximum number of iterations = KBRIEF output code K
RSMOOTH C, put rough into C, smooth into CSMOOTH by 3RSSH, twice
CPLOT condensed plot of y in C vs x in CLINES = KCHARACTERS = KXBOUNDS = from K to KYBOUNDS = from K to K
CTABLE coded table data in C, row levels in C, column levels in CMAXIMUM value in each cell should be codedMINIMUM value in each cell should be coded
ROOTOGRAM [data in C [use bin boundaries in C]]BOUNDARIES put in CDRRS put in CFITTED values, put in C
FITCREPLICATES K…KPOWER K…KSREPLICATES CSEFFECT CSPOWER CSCPBLOCK C
Quality Planning Tools%RRUN draw run chart with tests for randomness based on runs
RSUB data in rows of C…C (each row is a subgroup)CSUB data in C, subgroup size is EMEANS plot subgroup means (default)MEDIANS plot subgroup mediansTITLE for graph is "text"GSAVE save the graph in "filename"
%PARETO generate Pareto chart for COTHERS group cumulative percent over K into "Others"COUNTS are in CBY CSEPARATE chartsINDEPEND charts
TITLE for graph is "text"GSAVE save the graph in "filename"
%FISHBONE generate fishbone (cause-and-effect) diagramBR1 branch 1 list of causes in CBR2 branch 2 list of causes in CBR3 branch 3 list of causes in CBR4 branch 4 list of causes in CBR5 branch 5 list of causes in CBR6 branch 6 list of causes in CNOEMPTY suppress empty branchesNAM1 branch 1 name is "text"NAM2 branch 2 name is "text"NAM3 branch 3 name is "text"NAM4 branch 4 name is "text"NAM5 branch 5 name is "text"NAM6 branch 6 name is "text"NOLABELS suppress default branch labelsEFFECT name of effect is "text"TITLE for diagram is "text"GSAVE save the diagram in "filename"
%MULTIVAR Multi-Vari chartRESPONSE CFACTORS C…CDATAON display individual observations, just means
MU1ON turn on connect lines between means for factor 1MU2ON turn on connect lines between means for factor 2MU3ON turn on connect lines between means for factor 3TITLE 'string'
%SYMPLOT symmetry plot of data in C…CTITLE 'string'
Measurement Systems Analysis%GAGEAOV gage R&R using ANOVA method
PARTS in COPERS in CRESPONSE in CSTUDYVAR is KSIGMA is KTOLER is KNOPCONTRIBUTION no percent contribution outputNOPSTUDYVAR no percent study var. outputSEPARATE graphsTITLE for graph is "text"GAGE name is "text"DATE is "text"USER is "text"
%GAGEXBR gage R&R analysis using Average and Range MethodPARTS in COPERS in CRESPONSE in CSTUDYVAR is KSIGMA is KTOLER is KNOPCONTRIBUTION no percent contribution outputNOPSTUDYVAR no percent study var. outputSEPARATE graphsTITLE for graphs is "text"GAGE name is "text"DATE is "text"USER is "text"GTOL is "text"MISC comments are "text"
%GAGERUN plot response values by operator and part numberPARTS in COPERS in CTRIAL numbers in CRESPONSE in CMU is K
TITLE for chart is "text"GAGE name is "text"DATE is "text"USER is "text"GTOL is "text"MISC comments are "text"
%GAGENEST nested gage R&RPARTS in COPERS in CRESPONSE in CSTUDYVAR multiplier = KSEPARATE page for graphsSIGMA = KTOLER = KGAGE K (for on graph(s))DATE K (for on graph(s))USER K (for on graph(s))MISC K (for on graph(s))GTOL K (for on graph(s))TITLE for graph is K or "text string"NOPCONTRIBUTION no percent contribution outputNOPSTUDYVAR no percent study var. output
%GAGELIN gage linearity and accuracy studyPARTS in C
MASTER in CRESPONSE in CSIGMA is KTITLE for graphs is "text"GAGE name is "text"DATE is "text"USER is "text"GTOL is "text"MISC comments are "text"
Variables Control ChartsDEFTEST for test number K, test element is K, …, K K%BOXCOX transformation for non-normal data
RSUB data in rows of C…C (each row is a subgroup)CSUB data in C, subgroup size is ELAMBDA is KSTORE transformed data in C (with CSUB)STORE transformed data in C…C (with RSUB)
XBARCHART [data in C, subgroup size is E]RSUB data in rows of C…C (each row is a subgroup)MU = KSIGMA = K
RBAR base σ estimate on average of subgroup rangesSBAR base σ estimate on average of subgroup standard deviationsMMR base σ estimate on median of the moving rangeRSPAN length of moving range = K (default = 2)ESTIMATE parameters based on subgroups K…K or CBOXCOX KHPROCESS historical stages in process data C
VALUES K…K or COMIT K…K or C from parameter estimate calculationsTEST use tests K…K or C (all 8 tests)SUBGROUP size = K to calculate control limits and center lineSTAMP CSLIMITS [K…K]
UPPER KLOWER K
Graphics Subcommands for Variables Control Charts (see page 146)
RCHART [data in C, subgroup size is E]RSUB data in rows of C…C (each row is a subgroup)SIGMA = KRBAR base σ estimate on average of subgroup rangesESTIMATE parameters based on subgroups K…K or CBOXCOX KHPROCESS historical stages in process data C
VALUES K…K or COMIT K…K or C from parameter estimate calculationsTEST use tests K…K or C (first 4 tests only)SUBGROUP size = K to calculate control limits and center lineSTAMP CSLIMITS [K…K]
UPPER KLOWER K
Graphics Subcommands for Variables Control Charts (see page 146)
SCHART [data in C, subgroup size is E]RSUB data in rows of C…C (each row is a subgroup)SIGMA = KSBAR base σ estimate on average of subgroup st.devs.ESTIMATE parameters based on subgroups K…K or CBOXCOX KHPROCESS historical stages in process data C
VALUES K…K or COMIT K…K or C from parameter estimate calculationsTEST use tests K…K or C (first 4 tests only)SUBGROUP size = K to calculate control limits and center lineSTAMP CSLIMITS [K…K]
LOWER KGraphics Subcommands for Variables Control Charts (see page 146)
%RXBARRRSUB data in rows of C…C (each row is a subgroup)CSUB data in C, subgroup size is EMU = KSIGMA = KRBAR base σ estimate on average of subgroup rangesBOXCOX KHPROCESS historical stages in process data C
VALUES K…K or COMIT K…K from parameter estimate calculationsTEST use tests K…K (all 8 tests)SUBGROUP size = K to calculate control limits and center lineSTAMP CTITLE for graph is "text"GSAVE save graph in "filename"
%RXBARSRSUB data in rows of C…C (each row is a subgroup)CSUB data in C, subgroup size is EMU = KSIGMA = KSBAR base σ estimate on average of subgroup st.devs.BOXCOX K
HPROCESS historical stages in process data CVALUES K…K or C
OMIT K…K from parameter estimate calculationsTEST use tests K…K (all 8 tests)SUBGROUP size = K to calculate control limits and center lineSTAMP CTITLE for graph is "text"GSAVE save graph in "filename"
ICHART data in CMU = KSIGMA = KMMR base σ estimate on median of the moving rangeRSPAN length of moving range = K (default = 2)ESTIMATE parameters based on subgroups K…K or CBOXCOX KHPROCESS historical stages in process data C
VALUES K…K or COMIT K…K or C from parameter estimate calculationsTEST use tests K…K or C (all 8 tests)STAMP CSLIMITS [K…K]
Graphics Subcommands for Variables Control Charts (see page 146)
MRCHART data in CSIGMA = KMMR base σ estimate on median of the moving rangeRSPAN length of moving range = K (default = 2)ESTIMATE parameters based on subgroups K…K or CBOXCOX KHPROCESS historical stages in process data C
VALUES K…K or COMIT K…K or C from parameter estimate calculationsTEST use tests K…K (first 4 tests only)STAMP CSLIMITS [K…K]
UPPER KLOWER K
Graphics Subcommands for Variables Control Charts (see page 146)%IMRCHART C
MU = KSIGMA = KMMR base σ estimate on median of the moving rangeRSPAN length of moving range = K (default = 2)BOXCOX KHPROCESS historical stages in process data C
VALUES K…K or COMIT K…K or C from parameter estimate calculationsTEST use tests K…K (all 8 tests)STAMP CTITLE for graph is "text"GSAVE save graph in "filename"
EWMACHART [data in C, subgroup size is E]RSUB data in rows of C…C (each row is a subgroup)MU = KSIGMA = KRBAR base σ estimate on average of subgroup rangesSBAR base σ estimate on average of subgroup st.devs.MMR base σ estimate on median of the moving rangeRSPAN length of moving range = K (default = 2)ESTIMATE parameters based on subgroups K…K or CBOXCOX KOMIT K…K or C from parameter estimate calculationsSUBGROUP size = K to calculate control limits and center lineWEIGHT = K in EWMA (default = 0.2)STAMP CSLIMITS [K…K]
Graphics Subcommands for Variables Control Charts (see page 146)
MACHART [data in C, subgroup size is E]RSUB data in rows of C…C (each row is a subgroup)MU = KSIGMA = KRBAR base σ estimate on average of subgroup rangesSBAR base σ estimate on average of subgroup st.devs.MMR base σ estimate on median of the moving rangeRSPAN length of moving range = K (default = 2)ESTIMATE parameters based on subgroups K…K or CBOXCOX KOMIT K…K or C from parameter estimate calculationsSUBGROUP size = K to calculate control limits and center lineSPAN length of moving averages = K (default = 3)STAMP CSLIMITS [K…K]
UPPER KLOWER K
Graphics Subcommands for Variables Control Charts (see page 146)%RCUSUM
RSUB data in rows of C…C (each row is a subgroup)CSUB data in C, subgroup size is ESIGMA = K
RBAR base σ estimate on average of subgroup rangesSBAR base σ estimate on average of subgroup st.devs.MMR base σ estimate on median of the moving rangeRSPAN length of moving range = K (default = 2)OMIT K…K from parameter estimate calculationsTARGET specification = K (default = 0)FIR initialize CUSUMs at K σ’s above and below 0RESET CUSUMs to initial values when out-of-controlPLAN h = K k = K (defaults: h = 4 k = 0.5)VMASK center V-mask on subgroup KTITLE for graph is "text"GSAVE save graph in "filename"
%RZONERSUB data in rows of C…C (each row is a subgroup)CSUB data in C, subgroup size is EMU = KSIGMA = KRBAR base σ estimate on average of subgroup rangesSBAR base σ estimate on average of subgroup st.devs.MMR base σ estimate on median of the moving rangeRSPAN length of moving range = K (default = 2)OMIT K…K from parameter estimate calculationsWEIGHT K K K K (defaults = 0 2 4 8)RESET cumulative score to zero when out of control
SCORES put cumulative zone scores in CDISPLAY K standardized observations (default = 25)TITLE for graph is "text"GSAVE save graph in "filename"
%ZMRCHART data in C, part/product name/number in CCENTER using data in CSIGMAS in C to standardize dataSEPARATE σ estimate for each run independently (default)POOLED σ estimateEQUAL pool all data (across runs and parts) to estimate σPROP method to estimate σDISPLAY K standardized observations (default = 25)STAMP CTITLE for graph is "text"GSAVE save graph in "filename"
%WBCHART 3-way chart using both between-subgroup and within-subgroupvariations
RSUBS data in rows of C…C (each row is a subgroup)MU CRSPAN CMMR base σ estimate on median of the moving rangeRBAR base σ estimate on average of subgroup rangesHPROCESS historical values in process data C
SLIMIT CTEST use tests K…K (all 8 tests) OMIT K from parameter estimate calculationsBOXCOX CSTAMP CTITLE for graph is "text"GSAVE save graph in "filename"
Attributes Control ChartsPCHART data in C, subgroup size is E
HPROCESS historical values in process data CVALUES K…K
OMIT K…K or C from parameter estimate calculationsESTIMATE parameters based on subgroups K…K or CP process proportion defective = KSUBGROUP size = K to calculate control limits and center lineTEST use tests K…K or CGraphics Subcommands for Attributes Control Charts (see page 146)
NPCHART C EHPROCESS historical values in process data C
VALUES K…KOMIT K…K or C from parameter estimate calculations
ESTIMATE parameters based on subgroups K…K or CP process proportion defective = KSUBGROUP size = K to calculate control limits and center lineTEST use tests K…K or CGraphics Subcommands for Attributes Control Charts (see page 146)
CCHART CHPROCESS historical values in process data C
VALUES K…KOMIT K…K or C from parameter estimate calculationsESTIMATE parameters based on subgroups K…K or CMU = KTEST use tests K…K or CGraphics Subcommands for Attributes Control Charts (see page 146)
UCHART C EHPROCESS historical values in process data C
VALUES K…KOMIT K…K or C from parameter estimate calculationsESTIMATE parameters based on subgroups K…K or CMU = KSUBGROUP size = K to calculate control limits and center lineTEST use tests K…K or CGraphics Subcommands for Attributes Control Charts (see page 146)
TEXT K K "text"all of its subcommands (see page 53)
LINE K K … K K or C Call of its subcommands (see page 53)
POLYGON K K … K K or C Call of its subcommands (see page 53)
MARKER K K … K K or C Call of its subcommands (see page 53)
Frame Customization Subcommands:
AXIS K [K K] or C [K K]all of its subcommands (see page 56)
TICK K [K…K or C] or C [K…K or C]all of its subcommands (see page 56)
GRID K or Call of its subcommands (see page 56)
REFERENCE K K…K or K C or C K…K or C Call of its subcommands (see page 56)
MINIMUM K KMAXIMUM K KSCFRAME include ticks and reference lines in scale calculationSCANNOTATION include markers, lines, and polygons in scale calculationSTAMP C
Character Control ChartsThese are character (typewriter-style, or standard) graphics commands. From the Session window or the Command Line Editor, type the command GSTD to enable these commands. To re-enable high-resolution (professional) graphics commands, type GPRO. These charts are available as high-resolution graphics commands with the same names and with many more options. Character control charts are not available via the menus.
GSTD enable character (standard) graphicsGPRO enable high-resolution (professional) graphics
XBARCHART for C…C, subgroups are in EMU = KSIGMA = KRBAR use R to estimate σRSPAN = KTEST K…KSUBGROUP size is KSLIMITS are K…KHLINES at E…EESTIMATE using just samples K…KYLABEL = 'text'XLABEL = 'text'YINCREMENT = K
Note Single quotes must be used around text strings with these subcommands.
YSTART at K [end at K]XSTART at K [end at K]TITLE = 'text'FOOTNOTE = 'text'
RCHART for C…C, subgroups are in ESIGMA = KRBAR use R to estimate σSUBGROUP size is KSLIMITS are K…KHLINES at E…EESTIMATE using just samples K…KYLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XSTART at K [end at K]TITLE = 'text'FOOTNOTE = 'text'
SCHART for C…C, subgroups are in ESIGMA = KRBAR use R to estimate σSUBGROUP size is KSLIMITS are K…K
Note Single quotes must be used around text strings with these subcommands.
HLINES at E…EESTIMATE using just samples K…KYLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XSTART at K [end at K]TITLE = 'text'FOOTNOTE = 'text'
ICHART for C…CMU = KSIGMA = KRSPAN = KTEST K…KSLIMITS are K…KHLINES at E…EESTIMATE using just samples K…KYLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XSTART at K [end at K]TITLE = 'text'
Note Single quotes must be used around text strings with these subcommands.
Note Single quotes must be used around text strings with these subcommands.
MACHART for C…C, subgroups are in EMU = KSIGMA = KRBAR use R to estimate σSPAN = KRSPAN = KSUBGROUP size is KSLIMITS are K…KHLINES at E…EESTIMATE using just samples K…KYLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XSTART at K [end at K]TITLE = 'text'FOOTNOTE = 'text'
EWMACHART for C…C, subgroups are in EMU = KSIGMA = KRBAR use R to estimate σWEIGHT = K
Note Single quotes must be used around text strings with these subcommands.
RSPAN = KSUBGROUP size is KSLIMITS are K…KHLINES at E…EESTIMATE using just samples K…KYLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XSTART at K [end at K]TITLE = 'text'FOOTNOTE = 'text'
MRCHART for C…CSIGMA = KRSPAN = KSLIMITS are K…KHLINES at E…EESTIMATE using just samples K…KYLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XSTART at K [end at K]
Note Single quotes must be used around text strings with these subcommands.
Note Single quotes must be used around text strings with these subcommands.
PCHART number of nonconformities are in C…C, sample size = EP = KTEST K…KSUBGROUP size is KSLIMITS are K…KHLINES at E…EESTIMATE using just samples K…KYLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XSTART at K [end at K]TITLE = 'text'FOOTNOTE = 'text'
NPCHART number of nonconformities are in C…C, sample size = EP = KTEST K…KSUBGROUP size is KSLIMITS are K…KHLINES at E…EESTIMATE using just samples K…K
Note Single quotes must be used around text strings with these subcommands.
YLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XSTART at K [end at K]TITLE = 'text'FOOTNOTE = 'text'
CCHART number of nonconformities are in C…CMU = KTEST K…KSLIMITS are K…KHLINES at E…EESTIMATE using just samples K…KYLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XSTART at K [end at K]TITLE = 'text'FOOTNOTE = 'text'
UCHART number of nonconformities are in C…C, sample size = EMU = KTEST K…K
Note Single quotes must be used around text strings with these subcommands.
Note Single quotes must be used around text strings with these subcommands.
SUBGROUP size is KSLIMITS are K…KHLINES at E…EESTIMATE using just samples K…KYLABEL = 'text'XLABEL = 'text'YINCREMENT = KYSTART at K [end at K]XSTART at K [end at K]TITLE = 'text'FOOTNOTE = 'text'
Process Capability%RCAPA
RSUB data in rows of C…C (each row is a subgroup)CSUB data in C, subgroup size is EWITHIN use within-subgroup variation (Cp, Cpk)BETWEEN use between-subgroup variation (Pp, Ppk)MU = KSIGMA = KRBAR base σ estimate on average of subgroup rangesSBAR base σ estimate on average of subgroup st.devs.
Note Single quotes must be used around text strings with these subcommands.
MMR base σ estimate on median of the moving rangeRSPAN length of moving range = K (default = 2)TOTAL estimate σ using overall standard deviationMINMAX min = K max =K for capability histogramPERCENT in place of parts-per-millionSTORE in C…C statistics specified on HOARDHOARD store K…K (see arguments below) in cols. specified on STORE
BIASED do not use unbiasing constantsSRMSSD base σ estimate on the square root of half of the mean of the
squared successive differences when all subgroup sizes = 1ZBENCH replace capability statistics with benchmark Z’sBOXCOX CUSPEC K [K] (second K is flag for “hard” specification)LSPEC K [K] (second K is flag for “hard” specification)TARGET specification = KTOLER = K (default = 6)TITLE for graph is "text"GSAVE save graph in "filename"
USPEC and LSPEC are the specification limits. You must use at least one of these.
%BWCAPA capability analysis for between/withinCSUBS data in C…C, subgroup size is ERSUBS data in rows of C…C (each row is a subgroup)USPEC K [K] (second K is flag for “hard” specification)LSPEC K [K] (second K is flag for “hard” specification)TARGET specification = KMU = KBSIG = KWSIG = KBOXCOX KTOLER = K (default = 6)RBAR base σ estimate on average of subgroup rangesSBAR base σ estimate on average of subgroup st.devs.MMR base σ estimate on median of the moving rangeSRMSSD base σ estimate on the square root of half of the mean of the
squared successive differences when all subgroup sizes = 1RSPAN length of moving range = K (default = 2)TITLE for graph is "text string"GSAVE save graph in "filename"WITHIN use within-subgroup variation (Cp, Cpk)OVERALL use between-subgroup variation (Pp, Ppk)MINMAX capability histogram minimum = K maximum = KPERCENT in place of parts-per-millionBIASED do not use unbiasing constantsHOARD store K…K (see arguments on page 157) in cols. specified on STORESTORE in C…C statistics specified on HOARDNOCHART do not display graph
%WCAPA capability report, assuming data from Weibull distributionCSUBS data in CRSUBS data in rows of C…C USPEC K [K] (second K is flag for "hard" specification)LSPEC K [K] (second K is flag for "hard" specification)TARGET specification = KSHAPE KSCALE KTOLER = KTITLE for graph is K or "text"GSAVE save graph in "filename"
RSUB data in rows of C…C (each row is a subgroup)CSUB data in C, subgroup size is EMU = KSIGMA = KRBAR base σ estimate on average of subgroup rangesSBAR base σ estimate on average of subgroup st.devs.MMR base σ estimate on median of the moving rangeRSPAN length of moving range = K (default = 2)TOTAL estimate σ using overall standard deviationBIASED do not use unbiasing constantsSRMSSD base σ estimate on the square root of half of the mean of the
squared successive differences when all subgroup sizes = 1TARGET = KBOXCOX CUSPEC KLSPEC KTOLER = K (default = 6)BREAKOUT K subgroups or observations to display in run chartTEST use tests K…KTITLE for graph is "text"GSAVE save graph in "filename"
%BWSIXPAC capability sixpack for between/withinCSUBS data in C, subgroup size is ERSUBS data in rows of C…C (each row is a subgroup)USPEC KLSPEC KTARGET KMU = KBSIG KWSIGBOXCOX KTOLER = KRBAR base σ estimate on average of subgroup rangesSBAR base σ estimate on average of subgroup st.devs.MMR base σ estimate on median of the moving rangeSRMSSD base σ estimate on the square root of half of the mean of the
squared successive differences when all subgroup sizes = 1RSPAN length of moving range = K (default = 2)TITLE for graph is "text string"GSAVE save graph in "filename"BIASED do not use unbiasing constantsTEST K…K
%WSIXPACKCSUBS data in C, subgroup size is KRSUBS data in rows of C…C (each row is a subgroup)
USPEC KLSPEC KSHAPE KSCALE KTOLER = KBREAKOUT K subgroups or observations to display in run chartTITLE for graph is "text"GSAVE save graph in "filename"
%BCAPA capability report, assuming data from binomial distributionDEFECTS CSAMPLEN C or KP KTEST K…KTARGET specification = KBWFCOLORTITLE for graph is "text"
%PCAPA capability report, assuming data from Poisson distributionDEFECTS CSAMPLEN C or KMU = KTEST K…KTARGET specification = KBW
MKM plot positions determined by modified Kaplan-Meier methodKM plot positions determined by Kaplan-Meier methodHJ plot positions determined by Herd-Johnson methodALLPT plot all points when there are tied failure timesAVEPT plot average (median) of tied points when there are tied failure
timesMAXPT plot maximum of tied points when there are tied failure times
Censoring Options
CENSOR indicators are in C…CCVALUE censoring values KUPTIME upper limit for failure time is KUPFAIL upper limit for number of failures is K
KM Kaplan-Meier methodINTBY specify intervals from 0 to K1 by K2 (Actuarial method)MLE maximum likelihoodLSXY least squares (failure time (X) versus rank (Y))
Distributions
WEIBULL distributionEXTVALUE (Extreme Value) distributionEXPONENTIAL distributionLOGISTIC distributionLLOGISTIC (loglogistic) distributionNORMAL distributionLGNEXP (lognormal base EXP) distributionLGNTEN (lognormal base TEN) distribution
TITLE for graph is "text"GSAVE save graph in "file"
TURNBULL plot positions are determined by Turnbull methodACTUARIAL plot positions are determined by Actuarial method
Estimation Options
MLE maximum likelihoodLSXY least squares (failure time (X) versus rank (Y))
Distributions
NORMAL distributionWEIBULL distributionLGNEXP (lognormal base EXP) distributionLGNTEN (lognormal base TEN) distributionEXPO exponential distributionEXTVALUE (Extreme Value) distributionLOGISTIC distributionLLOGISTIC (loglogistic) distribution
TITLE for graph is "text"GSAVE save graph in "file"
%LDIDENT C…C (changed subcommands) FREQUENCY are in C…C
MKM plot positions determined by modified Kaplan-Meier methodKM plot positions determined by Kaplan-Meier methodHJ plot positions determined by Herd-Johnson methodALLPT plot all points when there are tied failure timesAVEPT plot average (median) of tied points when there are tied failure
timesMAXPT plot maximum of tied points when there are tied failure times
Censoring Options
CENSOR indicators are in C…CCVALUE censoring values KUPTIME upper limit for failure time is KUPFAIL upper limit for number of failures is K
Estimation Options
MLE maximum likelihoodLSXY least squares (failure time (X) versus rank (Y))PTILES percentiles for K…K
PTILES K…KMLE maximum likelihoodLSXY least squares (failure time (X) versus rank (Y))
Probability Plot Options
TURNBULL plot positions are determined by Turnbull methodACTUARIAL plot positions are determined by Actuarial method
Distributions
NORMAL distributionWEIBULL distributionLGNEXP (lognormal base EXP) distributionLGNTEN (lognormal base TEN) distributionEXPO exponential distributionEXTVALUE (Extreme Value) distributionLOGISTIC distributionLLOGISTIC (loglogistic) distribution
TITLE for graph is "text"GSAVE save graph in "file"
LTEST C…CFREQUENCY are in C…CBY CCOMSHAPE assume common shape parameterCOMSCALE assume common scale parameterLOWCI lower confidence bandUPCI upper confidence bandTWOCI two-sided confidence intervalTVSS test shape or scale = KTVSL test scale or location = KTESS test for equal shape or scaleTESL test for equal scale or location
Graph Options:
XMINIMUM x-axis scale minimum = KXMAXIMUM x-axis scale maximum = KXLABEL x-axis label = K or "text string"PPLOT does a probability plot
MKM plot positions determined by mod. Kaplan-Meier methodKM plot positions determined by Kaplan-Meier methodHJ plot positions determined by Herd-Johnson methodALLPT plot all points when there are tied failure timesAVEPT plot average (median) of tied points when there are tied
MAXPT plot maximum of tied points when there are tied failure times
CI display confidence intervals on probability and survival plotSPLOT does a survival plotHPLOT does a hazard plot
Censoring Options
CENSOR indicators are in C…C(if you have multiply right censored data)CVALUE censoring value KUPTIME value is K (if you have Type I censoring on the right)UPFAIL value is K (if you have Type II censoring on the right)
Distributions
WEIBULL distributionEXTVALUE (extreme value) distributionEXPONENTIAL distributionLOGISTIC distributionLLOGISTIC (loglogistic) distributionNORMAL distributionLGNEXP (lognormal base EXP) distributionLGNTEN (lognormal base TEN) distribution
MLE maximum likelihoodLSXY least squares (failure time (X) versus rank (Y))TIMES are K…K or C
STIME put in C SURVPROBABILITY put in C…C CLSURVPROBABILITY put in C…C
PTILES are K…K or CKM Kaplan-Meier method
SURVPROBABILITY put in C…CSTIME put in C…CSESURVPROBABILITY put in C…CCLSURVPROBABILITY put in C…CHAZARD put in C…CHTIME put in C…C
INTBY specify intervals from 0 to K1 by K2 (Actuarial method)SURVPROBABILITY put in C…CSTIME put in C…CSESURVPROBABILITY put in C…CCLSURVPROBABILITY put in C…CHAZARD put in C…CHTIME put in C…C
COMSHAPE assume common shape parameterCOMSCALE assume common scale parameterLOWCI lower confidence bandUPCI upper confidence bandTWOCI two-sided confidence intervalTVSS test shape or scale = KTVSL test scale or location = KTESS test for equal shape or scaleTESL test for equal scale or location
Graph Options
XMINIMUM x-axis minimum = KXMAXIMUM x-axis maximum = KXLABEL x-axis label = K "text string"PPLOT
TURNBULL plot positions are determined by Turnbull methodACTUARIAL plot positions are determined by Actuarial method
SPLOT does a survival plotHPLOT does a hazard plotCI display confidence intervals on probability and survival plot
Distributions
WEIBULL distributionEXTVALUE (extreme value) distributionEXPONENTIAL distribution
LOGISTIC distributionLLOGISTIC (loglogistic) distributionNORMAL distributionLGNEXP (lognormal base EXP) distributionLGNTEN (lognormal base TEN) distribution
Estimation Options
MLE maximum likelihoodLSXY least squares (failure time (X) versus rank (Y))TIMES are K…K or C
STIME put in C SURVPROBABILITY put in C…C CLSURVPROBABILITY put in C…C
PTILES are K…K or CTURNBULL calculate the Turnbull estimates
SURVPROBABILITY put in C…CSTIME put in C…CSESURVPROBABILITY put in C…CCLSURVPROBABILITY put in C…C
ACTUARIAL calculate the actuarial estimatesSURVPROBABILITY put in C…CSTIME put in C…CSESURVPROBABILITY put in C…CCLSURVPROBABILITY put in C…C
BRIEF output code KSTEP display log-likelihood for each iteration
Regression with Life DataLREGRESSION C…C = model
FREQUENCY are in C…C
Graph Options
RPLOT does a relation plotPTILES are in K…K or CPFAILURES plot uncensored values or midpoint of interval for interval
censored valuesDVALUE include design value K on relation plotALLCI include c.i.’s on all percentiles in relation plotONECI include c.i.’s on the middle percentile in relation plot
SPPLOT does a probability plot for the standardized residualsEPPLOT does an exponential probability plot for the Cox-Snell residualsCI include confidence intervals on probability plots
AFPPLOT probability plot based on fitted modelONECI place conf. intervals on design value fitted line if givenALLCI place confidence intervals on all fitted linesDVALUE enter design value to be placed on plot
AIPPLOT probability plot based on individual fits
Censoring Options
CENSOR indicators are in C…CCVALUE censoring value KSEND response variables are in an arbitrary censoring format
Model Options
REFERENCE level of C is K, …, C KFACTORS are in C…C
These subcommands only work with a model which has one covariate:
LOGISTIC distributionLLOGISTIC (loglogistic) distributionNORMAL distributionLGNEXP (lognormal base EXP) distributionLGNTEN (lognormal base TEN) distribution
Estimation Options
ESTIMATE percentiles or survival probabilities for [C…C] or [K…K] or for the predictor values in the data set [No arguments]
TIMES are in K…K or CPTILES are in K…K or CPERCENTILE put in C…CSEPERCENTILE put in C…CCLPERCENTILE put in C…CSURVPROBABILITY put in C…CCLSURVPROBABILITY put in C…C
CONFIDENCE level is K
Storage of Residuals
RESIDUAL put in C…CSRESIDUAL put standardized residuals in C…CCSREDISDUAL put Cox-Snell residuals in C…C
SF response is in a success failure formatFT response is in a failure trial format
Graph Options
PPLOT does a probability plotSPLOT does a survival plotCI display confidence intervals on probability and survival plotPPEARSON plot Pearson residuals versus probabilityPDEVIANCE plot deviance residuals versus probability
Model Options
FACTOR is in CREFERENCE level or event of C is K, …,C K
Distributions
WEIBULL distributionEXTVALUE (extreme value) distributionLOGISTIC distributionLLOGISTIC (loglogistic) distributionNORMAL distributionLGNEXP (lognormal base EXP) distributionLGNTEN (lognormal base TEN) distribution
SSTRESS put in C SURVPROBABILITY put in C…C CLSURVPROBABILITY put in C…C
HOSMER number of groups is KNAPPROXIMATION normal approximation confidence intervalsCONFIDENCE level is K
Storage
PRESIDUAL put in CDRESIDUAL put in CPPERCENT put in CPERCENTILE put in C…C SEPERCENTILE put in C…C CLPERCENTILE put in C…C EPROBABILITY put in C [C]COEFFICIENT put in CSECOEFFICIENT put in CNATRESPONSE put in KSENATRESPONSE put in KVCCOEFFICIENT put in MLOGLIKELIHOOD put in K
ITERATION maximum number is KSTART values are in CSETC set natural response to KOPTC optimize natural response
Session Window Subcommand Only
CTOLERANCE convergence criteria is K K
Output Options
BRIEF output code KSTEP display log-likelihood for each iteration
Factorial DesignsFFDESIGN K factors [K runs]
BLOCKS number is K or termlistREPLICATES for each point are KCPBLOCK add K center points to each blockFRACTION use KFOLD design on [fact…fact]ADD to design fact = term … fact = termALIAS print interactions up to order KRANDOMIZE [using base K]
XMATRIX put in C…CLEVELS are in K K, …, K K or CTERMS put in CSORDER put in C CBRIEF output code K
FDESIGN K…KBLOCKREPS KLEVELS K…K … K…KRANDOMIZE [K]SORDER C CXMATRIX C…C
PBDESIGN K factors [K runs]REPLICATES are K for each pointCENTER add K pointsRANDOMIZE [using base K]XMATRIX put in C…CLEVELS are K K, …, K K or CSORDER put in C C
FFACTORIAL modelBLOCKS CINUNIT K (K = 0 if design is in coded form; 1 if uncoded form)LEVELS are K K, …, K K or C
CTPT CFITC fits the center point column as a term in the modelCOVARIATES are in C…CSMEANS store means C…CSSTD store standard deviations C…CEFFECTS put in C…CCOEFFICIENTS put in C…CFITS put in C…CRESIDUALS put in C…CSRESIDUALS put in C…CTRESIDUALS put in C…CHI put in CCOOKD put in C…CDFITS put in C…CXMATRIX put in MMEANS for termlistGEFFECTS draw normal plot of effectsGPARETO draw Pareto chart of effectsRTYPE plot residual type KGHISTOGRAM display histogram of residualsGNORMAL display normal plot of residualsGFITS plot residuals vs fitsGORDER plot residuals vs worksheet order [or vs order specified in C] GVARIABLE plot residuals vs C…C
ALIAS [information up to order K]TOLERANCE K [K]BRIEF output code K
%FFMAIN main effects plots for factors in C…CMIXTURE treat the factors as process variables in a mixture designRESPONSE data in C…CYMIN y-axis minimum(s) is K […K] YMAX y-axis maximum(s) is K […K] INUNIT K (0 if design is in coded units, 1 if uncoded units)CTPT column that identifies the center point rows is CLSMEANS are in C…C (cannot be used with MIXTURE subcommand) LEVELS for the uncoded units are K K, …, K KTITLE for graph is "text"
%FFINT interaction plots for factors in C…CMIXTURE treat the factors as process variables in a mixture designRESPONSE data in C…C FULL display full matrix of plotsYMIN y-axis minimum(s) is K […K] YMAX y-axis maximum(s) is K […K] INUNIT K (0 if design is in coded units, 1 if uncoded units)CTPT column that identifies the center point rows is CLSMEANS are in C…C (cannot be used with MIXTURE subcommand) LEVELS for the uncoded units are K K, …, K KTITLE for graph is "text"
%FFCUBE cube plot for factors in C…CMIXTURE treat the factors as process variables in a mixture designRESPONSE data in C…CINUNIT K (0 if design is in coded units, 1 if uncoded units)CTPT column that identifies the center point rows is CLSMEANS are in C…C (cannot be used with MIXTURE subcommand) LEVELS for the uncoded units are K K, …, K KTITLE for graph is "text"
%GFMAIN main effects plots for factors in C…CRESPONSE data in C…CYMIN y-axis minimum(s) K […K]YMAX y-axis maximum(s) K […K]LSMEANS are in C…CTITLE "text"
%GFINT interaction plots for factors in C…CRESPONSE data in C…CLSMEANS are in C…CYMIN y-axis minimum(s) K […K]YMAX y-axis maximum(s) K […K]FULL displays the full matrix of plotsTITLE "text"
BLOCKS block the designCENTER add K pointsRANDOMIZE [using base K]XMATRIX put in C…CLEVELS are in K K, …, K K or CSORDER put in C CBRIEF output code K
CCDESIGN K factorsHALF use a one-half fraction designALPHA = KBLOCKS total number is K
CENTER add K to cube [add K points to axial block]RANDOMIZE [using base K]XMATRIX put in C…CLEVELS are in K K, …, K K or CINSCRIBE SORDER put in C CBRIEF output code K
FITS put in C…CRESIDUALS put in C…CSRESIDUALS put in C…CTRESIDUALS put in C…CHI put in CCOOKD put in C…CDFITS put in C…CCOEFFICIENTS put in C…CXMATRIX put in MQUADRATIC put in C…CRTYPE plot residual type KGHISTOGRAM display histogram of residualsGNORMAL display normal plot of residualsGFITS plot residuals vs fitsGORDER plot residuals vs orderGVARIABLE plot residuals vs C…CBRIEF KTOLERANCE K [K]
SCFRAME include ticks and reference lines in scale calculationSCANNOTATION include markers, lines, and polygons in scale calculationFRAME (default)NOFRAME [K]
Regions Commands and Subcommands for RSCONTOUR:
LAYOUTENDLAYOUT
FIGURE [K K K K]DATA [K K K K]ASPECT K K
File I/O Commands and Subcommands for RSCONTOUR:
GSAVE save graph in "filename"REPLACENOREPLACE
GVIEW "filename.mgf"GPAUSE [K]
RSSURFACE C C (RS Wireframe menu command with different settings.)FACTORIAL
Mixture DesignsSCDESIGN simplex centroid design for K components
AUGMENT augment design with axial points halfway between each vertexand centroid
AXIAL augment design with axial points halfway between each vertexand centroid
ALLREP replicate all points in the design K timesREPS K K … K K in each pair first value is point type; second is
number of replicatesTOTAL mixture totals are K [K…K] or CLOWER lower bounds for components are K…K or CUPPER upper bounds for components are K…K or CPROC K process variables [fraction = K [fraction ID = K]]LEVELS low and high levels for each process variable are K…KRANDOM randomize order [using base = K]XMATRIX store design matrix in C…CSORDER store standard order in C, store run order in CPTTYPE store point type = C (0 = center point, 1 = vertex, 2 = edge
centroid, 3 = face centroid)STOTAL store one or more totals in CSLOWER store component lower bounds in CSUPPER store component upper bounds in CPSEUDO display and store design in L-pseudo componentsPROPORTION display and store design in proportionsBRIEF output code K
SLDESIGN simplex lattice design for K components, degree of lattice = KAUGMENT augment design with center point and axial points halfway
between each vertex and centroidDEGREE of lattice = KCENTER augment design with center pointAXIAL augment design with axial points halfway between each
vertex and centroidALLREP replicate all points in the design K timesREPS K K … K K in each pair first value is point type; second is
number of replicatesTOTAL mixture totals are K [K…K] or CLOWER lower bounds for components are K…K or CUPPER upper bounds for components are K…K or CPROC K process variables [fraction = K [fraction ID = K]]LEVELS low and high levels for each process variable are K…KRANDOM randomize order [using base = K]
XMATRIX store design matrix in C…CSORDER store standard order in C, store run order in CPTTYPE store point type = C (0 = center point, 1 = vertex, 2 = edge
centroid, 3 = face centroid)STOTAL store one or more totals in CSLOWER store component lower bounds in CSUPPER store component upper bounds in CPSEUDO display and store design in L-pseudo componentsPROPORTION display and store design in proportionsBRIEF output code K
EVDESIGN extreme vertices design for K componentsAUGMENT augment design with center point and axial points halfway
between each vertex and centroidCENTER augment design with center pointDEGREE of design = KAXIAL augment design with axial points halfway between each
vertex and centroidALLREP replicate all points in the design K timesREPS K K … K K in each pair first value is point type; second is
number of replicatesTOTAL mixture totals are K [K…K] or CLOWER lower bounds for components are K…K or CUPPER upper bounds for components are K…K or CLINEAR linear constraint is K…K or CPROC K process variables [fraction = K [fraction ID = K]]LEVELS low and high levels for each process variable are K…KRANDOM randomize order [using base = K]XMATRIX store design matrix in C…CSORDER store standard order in C, store run order in CPTTYPE store point type = C (0 = center point, 1 = vertex, 2 = edge
centroid, 3 = face centroid)STOTAL store one or more totals in CSLOWER store component lower bounds in CSUPPER store component upper bounds in CSLINEAR store linear constraints in C…CPSEUDO display and store design in L-pseudo componentsPROPORTION display and store design in proportionsBRIEF output code K
SIMPLEX simplex design plot, components of the design are in C…CTOTAL mixture totals are K [K…K] or CLOWER lower bound of components are K…K or CUPPER upper bound of components are K…K or CLINEAR linear constraint is K…K or CSINGLE plot with components C C CALLTRIPLETS multiple plots to show all component tripletsPROPORTION proportion unit for the component axes on the plotPSEUDO pseudocomponent unit for the component axes on the plotPROCESS process variables in C…CAMOUNT mixture amount variable in C LABEL design points with replicates [or with values in C]CLOW set values of all components not on plot to their lower boundCHIGH set values of all components not on plot to their upper boundCHOLD set all components not on plot to K…K
FLOW set levels of all process variables at low level (= −1)FHIGH set levels of all process variables at high level (= 1)FHOLD set levels of process variables K…KAHOLD set mixture amount at KMINIMUM value of simplex coordinate K (1, 2, or 3) is KGRID draw grid lines on each axis [at K intervals]TITLE for graph is "text"
MIXREG C…C = TERMSBLOCKS blocking variable in CPROC process variables in C…CAMOUNT mixture amount in CPSEUDO fit model in pseudocomponentsMODEL store model term codes in C…CSTEPWISE model selection procedureFORWARD model selection procedureBACKWARD model selection procedureFORCE TERMS to be forced in modelENTER TERMS in starting modelAENTER α = K to enter term in modelAREMOVE α = K to remove term from modelBEST K alternative termsFITS put in C…CRESIDUALS put in C…C
SRESIDUALS put in C…CTRESIDUALS put in C…CHI put in CCOOKD put in C…CDFITS put in C…CMODEL put mixture models in C…CCOEFFICIENTS put in C…CXMATRIX put in MRTYPE plot residual type KGHISTOGRAM plot histogram of residualsGNORMAL plot normal plot of residualsGFITS plot residuals vs fitsGORDER plot residuals vs orderGVARIABLE plot residuals vs variables in C…CTOLERANCE for collinear and constant predictors are K [K]BRIEF output code K
TRACE response trace plot with response in C, model in CTOTAL mixture total in K [K K] or CLOWER lower bound of components = K…K or CUPPER upper bound of components = K…K or CLINEAR linear constraint on components = K…K or CPIEPEL trace plot along Piepel directionsXREFERENCE K (row ID of reference blend)XREFERENCE K…K (coordinates of reference blend)FLOW set levels of all process variables to low (−1)FMIDDLE set levels of all process variables to average (0)FHIGH set levels of all process variables to high (1)FHOLD set process variables to levels K…KAHOLD set mixture amount to KCURVE
TYPE line type of trace curves = K…K (default = all solid)COLOR line color of trace curves = K…K (default = different colors)
NOHLEGEND do not display hold legend
Generic plot subcommands available
TICK draw ticks at K…KGRID KTEXT K K "text"TITLE for graph is "text"FOOT footnote for graph is "text"
DATA K K K KLEGEND K K K KNOLEGEND do not display curve legendMINIMUM minimum valueMAXIMUM maximum value of axis K is K
MIXSURFACE mixture surface plot with response in C, model in CTOTAL mixture total in K [K…K] or CLOWER lower bound of components = K…K or CUPPER upper bound of components = K…K or CLINEAR linear constraint on components = K…K or CSINGLE plot for components or process variables C C [C]ALLTRIPLETS multiple plots to show all component tripletsALLPAIRS multiple plots to show all numerical process variable pairsPROPORTION component axes displayed in proportion unitPSEUDO component axes displayed in pseudocomponent unitSURFACE
ECOLOR wireframe color = K (default = 1)COLOR surface color = K (default = 0)
NMESH number of mesh points in horizontal dir. is K, in vertical dir. is KCLOW set levels of all components not on plot at lower boundCMIDDLE set levels of all components not on plot at average of lower and
upper boundsCHIGH set levels of all components not on plot at upper boundCHOLD set levels of all components not on plot at K…KFLOW set levels of all process variables to low (−1)FMIDDLE set levels of all process variables to average (0)FHIGH set levels of all process variables to high (1)FHOLD set process variables to levels K…KAHOLD set mixture amount to KNOHLEGEND do not display hold legendMINIMUM minimum value of axis K is KMAXIMUM maximum value of axis K is KTITLE for graph is "text"WTITLE title for window is "text"TEXT K K "text"FOOT footnote for graph is "text"FIGURE K K K KDATA K K K KLEGEND K K K KNOLEGEND do not display legend
MIXCONTOUR mixture contour plot with response in C, model in CTOTAL mixture total in K [K…K] or C…CLOWER lower bound of components = K…K or C…CUPPER upper bound of components = K…K or C…C
LINEAR linear constraint on components = K…K or C…CSINGLE plot for components or process variables C C [C]ALLTRIPLETS multiple plots to show all component tripletsALLPAIRS multiple plots to show all numerical process variable pairsPROPORTION component axes displayed in proportion unitPSEUDO component axes displayed in pseudocomponent unitNLEVELS number of contour levels = KLEVELS values of contour levels = K…KNMESH number of mesh points in horizontal dir. is K, in vertical dir. is KCONNECT or AREA
TYPE of contour lines or areas = K…KCOLOR of contour lines or areas = K…K
CLOW set levels of all components not on plot at lower boundCMIDDLE set levels of all components not on plot at average of lower and
upper boundsCHIGH set levels of all components not on plot at upper boundCHOLD set levels of all components not on plot at K…KFLOW set levels of all process variables to low (−1)FMIDDLE set levels of all process variables to average (0)FHIGH set levels of all process variables to high (1)FHOLD set process variables to levels K…KAHOLD set mixture amount to KMINIMUM minimum value of axis K is KMAXIMUM maximum value of axis K is KGRID [place grid lines at intervals K]POINTS display design points on plotNOHLEGEND do not display hold legendTITLE for graph is "text"WTITLE title for window is "text"TEXT K K "text"FOOT footnote for graph is "text"FIGURE K K K KDATA K K K KLEGEND K K K KNOLEGEND do not display contour legend
OPTDES K optimal design points [design columns are in C…C]MIXMODEL TERMS (these are model term arguments)RESMODEL TERMS (these are model term arguments)AUGMENT [design specified by indicator column C]EVALUATE [design specified by indicator column C]BLOCK blocking variable in CPROCESS process variables in C…CAMOUNT mixture amount in CPSEUDO model fitted in pseudocomponent unitGENERATE K percent of initial design randomly [do K trials [use random
number seed K]]SEQUENTIAL method to generate initial designEXCHANGE K design point in improving the initial designFEDOROV method for improving the initial designINDICATOR store selected point indicator in CXMATRIX store optimal design matrix in C…CCARRY additional columns (C…C) from old worksheet to columns
QUADRATICS C…CMINACCEPT K…K or CMAXACCEPT K…K or CTARGET K…K or CGOAL K…K or CUTWEIGHTS K…K or CIMPORTANCE K…K or CSTART K…KFACTORIAL C…C
INUNIT K (0 = coded units, 1 = uncoded units)FLEVELS nominal factor levels are K K … K KCOVARIATES K…K
MIXTURE model C…CLINEAR K…KTOTAL amount of mixture components is CINUNIT K (0 = coded units, 1 = uncoded units)FLEVELS nominal factor levels are K K … K K
DSTORE COPTIPLOT
%OVERCONT overlaid contour plot for multiple responsesRESPONSES C…CQUADRATICS C…CSINGLE C CHOLD K…K
%FACTOVER overlaid contour plot for multiple responses of factorial designsRESPONSE variables C…CQUADRATICS in C…CINUNIT factor values = K, plot display = K (0 = coded, 1 = uncoded)
FLEVELS K…KSINGLE x-axis = C, y-axis = CHOLD extra factors and covariates at K…KLV1 contours for first response variable = K KLV2 contours for second response variable = K KLV3 contours for third response variable = K KLV4 contours for fourth response variable = K KLV5 contours for fifth response variable = K KLV6 contours for sixth response variable = K KLV7 contours for seventh response variable = K KLV8 contours for eighth response variable = K KLV9 contours for ninth response variable = K KLV10 contours for tenth response variable = K KYMIN y-axis minimum = KYMAX y-axis maximum = KXMIN x-axis minimum = KXMAX x-axis maximum = KCLEANUP remove model info. cols. from worksheet after plot is drawnTITLE for graph is "text"
RESPONSE in C…CMODEL in C…CMMERGE merge mixture model C…C
STORE in C…CLV1 contours = K KLV2 contours = K KLV3 contours = K KLV4 contours = K KLV5 contours = K KLV6 contours = K KLV7 contours = K KLV8 contours = K KLV9 contours = K KLV10 contours = K KTOTAL mixture amounts = K [K…K]LOWER bound of components = K…K UPPER bound of components = K…K LN1 linear constraint = K…KLN2 linear constraint = K…KLN3 linear constraint = K…KLN4 linear constraint = K…KLN5 linear constraint = K…KLN6 linear constraint = K…KLN7 linear constraint = K…K
LN8 linear constraint = K…KLN9 linear constraint = K…KLN10 linear constraint = K…KPSEUDO component scale for plotPROPORTION axis scale is in proportionSINGLE components or process variables C C [C] for 2D plotCHOLD components not on plot = K…KFHOLD extra process variables not on plot = K…KAHOLD mixture amount = KXMIN x-axis minimum = KYMIN y-axis minimum = KZMIN z-axis minimum = KXMAX x-axis maximum = KYMAX y-axis maximum = KCLEANUP remove model info. cols. from worksheet after plot is drawnTITLE for graph is "text"
Taguchi Designs%ROBUST analyze an inner-outer array design
RESPONSE data from the outer array are in C…CINNER array factors are in C…C
DYNAMIC signal in CYREF response reference value = KSREF signal reference value = KNOREF do not use reference values for the regression, estimate a
separate regression intercept for each runSLOPES put in CINTERCEPTS put in CLSTDEVS put loge of standard deviations in CGSLOPES generate plots for regression slopesINTMAT generate interaction plot matrix for C C … C CINTSEP generate a separate interaction plot for C C … C CTMEANS generate a response table for meansTSN generate a response table for signal-to-noise ratiosTSTDS generate a response table for standard deviationsTSLOPES generate a response table for slopesSCATTER generate scatter plots of data with fitted lines SNRATIO put s/n ratios in CMEANS put means in CSTDEVS put standard deviations in CCV put coefficients of variation in CNOM1 bases s/n ratio on standard deviations onlyNOM2 bases s/n ratio on means and st.devs. (default)LOWER s/n ratio models lower is betterHIGHER s/n ratio models higher is better
LOGSTD do a loge transformation on the standard deviationsGSN plot s/n ratiosGMEAN plot meansGSTD plot standard deviationsTITLE for plot is "text"
%PREDICTINNER array factors C…CCHARACTERISTICS to be predicted in C…CLABELS for predictions "text" … "text"LEVELS new levels where predictions are to be calculated are E…EINTERACTIONS C C … C CPFITS put predicted values in C…CUNITS inner array = K, units of LEVELS = K (0 = coded, 1 = uncoded)NLEVELS number of levels = K…KFLEVELS uncoded levels K…K
OADESIGN Taguchi orthogonal array design K runs with K levels for K factors[K levels for K factors]
XMATRIX put in C…CLEVELS K K, K K … K KCOLUMNS of the full orthogonal array to use = K … KINTERACTIONS to remain unconfounded are C C C C … C CSIGNAL factor levels are K…K put in C