Entity Tables, Relationship Tables We Classify using any Table (as the Training Table) on any of its columns, the class label column. Medical Expert System: Using entity Training Table Patients(PID,Symptom1,Symptom2,Symptom3 ,Disease) and class label Disease, we can classify new patients based on symptoms using Nearest Neighbor, Model based (Decision Tree, Neural Network, SVM, SVD, etc.) classification. Netflix Contest: Using relationship TrainingTable, Rents(UID,MID,Rating,Date) and class label Rating, classify new (UID,MID,Date) tuples. How (since there are no feature columns really)? Movie User Rents UID CNAME AGE TranID Rating Date MID Mname Date Netflix Item Customer buys C# CNAME AGE TranID Count Date I# Iname Suppl Date Price Market Basket Cours e Student Enrollments S# SNAME GEN C# S# GR C# CNAME ST TERM Educational Medical Ward Patients Has is in PID PNAME Symptom1 Symptom2 Symptom3 Disease WID Wname Capaci t y We Cluster on a Table also, but usually just to "prepare" a classification TrainingTable (move up a semantic hierarchy, e.g., in Items, cluster on PriceRanges in 100 dollar intervals or to determine classes in the first place (each cluster is then declared a separate class) ). We do Association Rule Mining (ARM) on relationships (e.g., the Market Basket Research; ARM on "buys" relationship). We can let Near Neighbor Movies vote? What makes a movie, mid, "near" to MID? If mid is rated similarly to MID by users, uid 1 , ..., uid .. Here we use correlations for near instead of an actual distance once again. We can let Near Neighbor Users vote? What makes a user, uid, "near" to UID? If uid rates similarly to UID on movies, mid 1 , ..., mid . So we use correlations for near instead of distance. This is typical when classifying on a relationship table.
Entity Tables, Relationship Tables We Classify using any Table (as the Training Table) on any of its columns, the class label column. Medical Expert System:
Dec 22, 2015
Welcome message from author
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.
Transcript
Entity Tables,Relationship Tables
We Classify using any Table (as the Training Table) on any of its columns, the class label column.
Medical Expert System: Using entity Training Table
Patients(PID,Symptom1,Symptom2,Symptom3,Disease) and class label Disease, we can classify new patients based on symptoms using Nearest Neighbor, Model based (Decision Tree, Neural Network, SVM, SVD, etc.) classification.
Netflix Contest: Using relationship TrainingTable,
Rents(UID,MID,Rating,Date) and class label Rating, classify new (UID,MID,Date) tuples.
How (since there are no feature columns really)?
Movie User RentsUID
CNAME
AGE
TranID
Rating
Date
MID
Mname
Date
Netflix
ItemCustomer buysC#
CNAME
AGE
TranID
Count
Date
I#
Iname
Suppl
Date
Price
Market Basket
CourseStudent EnrollmentsS#
SNAME
GEN
C#
S#
GR
C#
CNAME
ST
TERM
Educational
Medical
WardPatients Hasis inPID
PNAME
Symptom1
Symptom2
Symptom3
Disease
WID
Wname
Capacity
We Cluster on a Table also, but usually just to "prepare" a classification TrainingTable (move up a semantic hierarchy, e.g., in Items, cluster on PriceRanges in 100 dollar intervals or to determine classes in the first place (each cluster is then declared a separate class) ).
We do Association Rule Mining (ARM) on relationships (e.g., the Market Basket Research; ARM on "buys" relationship).
We can let Near Neighbor Movies vote? What makes a movie, mid, "near" to MID? If mid is rated similarly to MID by users, uid1, ..., uid..
Here we use correlations for near instead of an actual distance once again.
We can let Near Neighbor Users vote? What makes a user, uid, "near" to UID? If uid rates similarly to UID on movies, mid1, ..., mid.
So we use correlations for near instead of distance. This is typical when classifying on a relationship table.
Key a1 a2 a3 a4 a5 a6 a7 a8 a9 a10=C a11 a12 a13 a14 a15 a16 a17 a18 a19 a20
t12 1 0 1 0 0 0 1 1 0 1 0 1 1 0 1 1 0 0 0 1t13 1 0 1 0 0 0 1 1 0 1 0 1 0 0 1 0 0 0 1 1t15 1 0 1 0 0 0 1 1 0 1 0 1 0 1 0 0 1 1 0 0t16 1 0 1 0 0 0 1 1 0 1 1 0 1 0 1 0 0 0 1 0t21 0 1 1 0 1 1 0 0 0 1 1 0 1 0 0 0 1 1 0 1t27 0 1 1 0 1 1 0 0 0 1 0 0 1 1 0 0 1 1 0 0t31 0 1 0 0 1 0 0 0 1 1 1 0 1 0 0 0 1 1 0 1t32 0 1 0 0 1 0 0 0 1 1 0 1 1 0 1 1 0 0 0 1t33 0 1 0 0 1 0 0 0 1 1 0 1 0 0 1 0 0 0 1 1t35 0 1 0 0 1 0 0 0 1 1 0 1 0 1 0 0 1 1 0 0t51 0 1 0 1 0 0 1 1 0 0 1 0 1 0 0 0 1 1 0 1t53 0 1 0 1 0 0 1 1 0 0 0 1 0 0 1 0 0 0 1 1t55 0 1 0 1 0 0 1 1 0 0 0 1 0 1 0 0 1 1 0 0t57 0 1 0 1 0 0 1 1 0 0 0 0 1 1 0 0 1 1 0 0t61 1 0 1 0 1 0 0 0 1 0 1 0 1 0 0 0 1 1 0 1t72 0 0 1 1 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1t75 0 0 1 1 0 0 1 1 0 0 0 1 0 1 0 0 1 1 0 0
t12 0 0 1 0 1 1 0 2
t13 0 0 1 0 1 0 0 1
t15 0 0 1 0 1 0 1 2
a5 a6 a10=C a11 a12 a13 a14 distance
from 000000
Workspacefor the 3nearestneighborsso far
0 0 0 0 0 0
distance=2, don’t replace
0 0 0 0 0 0
distance=4, don’t replace
0 0 0 0 0 0
distance=4, don’t replace
0 0 0 0 0 0
distance=3, don’t replace
0 0 0 0 0 0
distance=3, don’t replace
0 0 0 0 0 0
distance=2, don’t replace
0 0 0 0 0 0
distance=3, don’t replace
0 0 0 0 0 0
distance=2, don’t replace
0 0 0 0 0 0
distance=1, replace
t53 0 0 0 0 1 0 0 1
0 0 0 0 0 0
distance=2, don’t replace
0 0 0 0 0 0
distance=2, don’t replace
0 0 0 0 0 0
distance=3, don’t replace
0 0 0 0 0 0
distance=2, don’t replace
0 1
C=1 wins!
WALK THRU a kNN classification example: 3NN CLASSIFICATION
of an unclassified sample, a=( a5 a6 a11 a12 a13 a14 ) = =( 0 0 0 0 0 0 ).
0 0 0 0 0 0
distance=2, don’t replace
Key a1 a2 a3 a4 a5 a6 a7 a8 a9 a10=C a11 a12 a13 a14 a15 a16 a17 a18 a19 a20
t12 1 0 1 0 0 0 1 1 0 1 0 1 1 0 1 1 0 0 0 1t13 1 0 1 0 0 0 1 1 0 1 0 1 0 0 1 0 0 0 1 1t15 1 0 1 0 0 0 1 1 0 1 0 1 0 1 0 0 1 1 0 0t16 1 0 1 0 0 0 1 1 0 1 1 0 1 0 1 0 0 0 1 0t21 0 1 1 0 1 1 0 0 0 1 1 0 1 0 0 0 1 1 0 1t27 0 1 1 0 1 1 0 0 0 1 0 0 1 1 0 0 1 1 0 0t31 0 1 0 0 1 0 0 0 1 1 1 0 1 0 0 0 1 1 0 1t32 0 1 0 0 1 0 0 0 1 1 0 1 1 0 1 1 0 0 0 1t33 0 1 0 0 1 0 0 0 1 1 0 1 0 0 1 0 0 0 1 1t35 0 1 0 0 1 0 0 0 1 1 0 1 0 1 0 0 1 1 0 0t51 0 1 0 1 0 0 1 1 0 0 1 0 1 0 0 0 1 1 0 1t53 0 1 0 1 0 0 1 1 0 0 0 1 0 0 1 0 0 0 1 1t55 0 1 0 1 0 0 1 1 0 0 0 1 0 1 0 0 1 1 0 0t57 0 1 0 1 0 0 1 1 0 0 0 0 1 1 0 0 1 1 0 0t61 1 0 1 0 1 0 0 0 1 0 1 0 1 0 0 0 1 1 0 1t72 0 0 1 1 0 0 1 1 0 0 0 1 1 0 1 1 0 0 0 1t75 0 0 1 1 0 0 1 1 0 0 0 1 0 1 0 0 1 1 0 0
WALK THRU of required 2nd scan to find Closed 3NN set. Does it change vote?
0 0 0 0 0 0
d=2, include it also
0 0 0 0 0 0
d=4, don’t include
0 0 0 0 0 0
d=4, don’t include
0 0 0 0 0 0
d=3, don’t include
0 0 0 0 0 0
d=3, don’t include
0 0 0 0 0 0
d=2, include it also
0 0 0 0 0 0
d=3, don’t include
0 0 0 0 0 0
d=2, include it also
0 0 0 0 0 0
d=1, already voted
0 0 0 0 0 0
d=2, include it also
0 0 0 0 0 0
d=2, include it also
0 0 0 0 0 0
d=3, don’t replace
0 0 0 0 0 0
d=2, include it also
0 0 0 0 0 0
d=2, already voted
0 0 0 0 0 0
d=1, already voted
0 1
Vote after 1st scan.YES! C=0 wins now!
t12 0 0 1 0 1 1 0 2
t13 0 0 1 0 1 0 0 1
a5 a6 a10=C a11 a12 a13 a14 distance
t53 0 0 0 0 1 0 0 1
Unclassified sample: 0 0 0 0 0 03NN set after 1st scan
0 0 0 0 0 0
d=2, include it also0 0 0 0 0 0
d=2, include it also
C
00000000001111111
C
11111111110000000
WALK THRU: C3NN using P-trees
a20
11001011101100110
keyt12
t13
t15
t16
t21
t27
t31
t32
t33
t35
t51
t53
t55
t57
t61 t72
t75
a1
11110000000000100
a2 00001111111111000
a3
11111100000000111
a4
000000000 01111011
a5
00001111110000100
a6
00001100000000000
a7
11110000001111011
a8
11110000001111011
a9
00000011110000100
C11111111110000000
a11
00011010001000100
a12 11100001110110011
a13 10011111001001110
a14 00100100010011001
a15 110100011001000 10
a16
10000001000000010
a17 00101110011011101
a18
00101110011011101
a19
01010000100100000
Ps
00000000000000000
a14
11011011101100110
a13
01100000110110001
a12
00011110001001100
a11
11100101110111011
a6
11110011111111111
a5
1111000000111 1011
No neighbors at distance=0
First let all training points at distance=0 vote, then distance=1, then distance=2, ... until 3 For distance=0 (exact matches) constructing the P-tree, Ps
then AND with PC and PC’ to compute the vote. (black denotes complement, red denotes uncomplemented
C
00000000001111111
C
11111111110000000
a20
11001011101100110
keyt12
t13
t15
t16
t21
t27
t31
t32
t33
t35
t51
t53
t55
t57
t61 t72
t75
a1
11110000000000100
a2 00001111111111000
a3
11111100000000111
a4
000000000 01111011
a5
00001111110000100
a6
00001100000000000
a7
11110000001111011
a8
11110000001111011
a9
00000011110000100
a10 =C11111111110000000
a11
00011010001000100
a12 11100001110110011
a13 10011111001001110
a14 00100100010011001
a15 110100011001000 10
a16
10000001000000010
a17 00101110011011101
a18
00101110011011101
a19
01010000100100000
PD(s,1)
01000000000100000
a14
11011011101100110
a13
01100000110110001
a12
00011110001001100
a11
11100101110111011
a6
11110011111111111
a5
00001111110000100
a14
11011011101100110
a13
01100000110110001
a12
00011110001001100
a11
11100101110111011
a6
0 0001100000000000
a5
1111000000111 1011
a14
11011011101100110
a13
01100000110110001
a12
00011110001001100
a11
00011010001000100
a6
11110011111111111
a5
1111000000111 1011
a14
11011011101100110
a13
01100000110110001
a12
1 1100001110110011
a11
11100101110111011
a6
11110011111111111
a5
1111000000111 1011
a14
11011011101100110
a13
1 0011111001001110
a12
00011110001001100
a11
11100101110111011
a6
11110011111111111
a5
1111000000111 1011
a14
0 0100100010011001
a13
01100000110110001
a12
00011110001001100
a11
11100101110111011
a6
11110011111111111
a5
1111000000111 1011
Construct Ptree, PS(s,1) = OR Pi = P|si-ti|=1; |sj-tj|=0, ji
= OR PS(si,1) S(sj,0)
OR
P5 P6 P11 P12 P13 P14
j{5,6,11,12,13,14}-{i}
0 1
i=5,6,11,12,13,14i=5,6,11,12,13,14
WALK THRU: C3NNCdistance=1 nbrs:
keyt12
t13
t15
t16
t21
t27
t31
t32
t33
t35
t51
t53
t55
t57
t61 t72
t75
a5
00001111110000100
a6
00001100000000000
a10 C11111111110000000
a11
00011010001000100
a12 11100001110110011
a13 10011111001001110
a14 00100100010011001
0 1
OR{all double-dim interval-Ptrees}; PD(s,2) = OR Pi,j
Pi,j = PS(si,1) S(sj,1) S(sk,0) k{5,6,11,12,13,14}-{i,j}
i,j{5,6,11,12,13,14}
a14
11011011101100110
a13
01100000110110001
a12
00011110001001100
a11
11100101110111011
a6
0 0001100000000000
a5
00001111110000100
a14
11011011101100110
a13
01100000110110001
a12
00011110001001100
a11
00011010001000100
a6
11110011111111111
a5
00001111110000100
a14
11011011101100110
a13
01100000110110001
a12
1 1100001110110011
a11
11100101110111011
a6
11110011111111111
a5
00001111110000100
a14
11011011101100110
a13
1 0011111001001110
a12
00011110001001100
a11
11100101110111011
a6
11110011111111111
a5
00001111110000100
a14
0 0100100010011001
a13
01100000110110001
a12
00011110001001100
a11
11100101110111011
a6
11110011111111111
a5
00001111110000100
P5,6 P5,11 P5,12 P5,13 P5,14
a14
11011011101100110
a13
01100000110110001
a12
00011110001001100
a11
00011010001000100
a6
0 0001100000000000
a5
1111000000111 1011
a14
11011011101100110
a13
01100000110110001
a12
11100001110110011
a11
11100101110111011
a6
0 0001100000000000
a5
1111000000111 1011
a14
11011011101100110
a13
11100001110110011
a12
00011110001001100
a11
11100101110111011
a6
0 0001100000000000
a5
1111000000111 1011
a14
00100100010011001
a13
01100000110110001
a12
00011110001001100
a11
11100101110111011
a6
0 0001100000000000
a5
1111000000111 1011
P6,11 P6,12 P6,13 P6,14
a14
11011011101100110
a13
01100000110110001
a12
11100001110110011
a11
00011010001000100
a6
11110011111111111
a5
1111000000111 1011
a14
11011011101100110
a13
10011111001001110
a12
00011110001001100
a11
00011010001000100
a6
11110011111111111
a5
1111000000111 1011
a14
11011011101100110
a13
01100000110110001
a12
00011110001001100
a11
00011010001000100
a6
11110011111111111
a5
1111000000111 1011
P11,12 P11,13 P11,14
a14
11011011101100110
a13
1 0011111001001110
a12
11100001110110011
a11
11100101110111011
a6
11110011111111111
a5
11110000001111011
a14
0 0100100010011001
a13
01100000110110001
a12
11100001110110011
a11
11100101110111011
a6
11110011111111111
a5
11110000001111011
P12,13 P12,14
We now have 3 nearest nbrs. We could quite and declare C=1 winner?
a14
0 0100100010011001
a13
10011111001001110
a12
00011110001001100
a11
11100101110111011
a6
11110011111111111
a5
11110000001111011
P13,14
We now have the C3NN set and we can declare C=0 the winner!
WALK THRU: C3NNC distance=2 nbrs:
In this example, there were no exact matches (dis=0 nbrs or similarity=6 neighbors) for the sample.
There were two nbrs found at a distance of 1 (dis=1 or sim=5) and nine dis=2, sim=4 nbrs.
All 11 neighbors got an equal votes even though the two sim=5 are much closer neighbors than the nine sim=4. Also processing for the 9 is costly.
A better approach would be to weight each vote by the similarity of the voter to the sample (We will use a vote weight function which is linear in the similarity (admittedly, a better choice would be a function which is Gaussian in the similarity, but, so far, it has been too hard to compute).
As long as we are weighting votes by similarity, we might as well also weight attributes by relevance also (assuming some attributes are more relevant than others. e.g., the relevance weight of a feature attribute could be the correlation of that attribute to the class label).
P-trees accommodate this method very well (in fact, a variation on this theme won the KDD-cup competition in 02 ( http://www.biostat.wisc.edu/~craven/kddcup/ ) and is published in the so-called Podium Classification methods.
Notice though that the Ptree method (Horizontal Processing of Vertical Data or HPVD) really relies on a bonafide distance, in this case Hamming Distance or L1. However, the Vertical Processing of Horizontal Data (VPHD) doesn't.
VPHD therefore works even when we use a correlation as the notion of "near".
If you have been involve in the DataSURG Netflix Prize efforts, you will note that we used Ptree technology to calculate correlations but not to find near neighbors through correlations.
Movie User RentsUID
CNAME
AGE
TID
Rating
Date
MID
Mname
Date
Again, Rents is also a Table ( Rents(TID, UID,MID,Rating,Date) ), using the class label column, Rating, we can classify potential new ratings. Using these predicted ratings, we can recommend Rating=5 rentals.
Does ARM require a binary relationship? YES! But since every Table gives a binary relationship between any two of its columns we can do ARM on them.
E.g., we could do ARM on Patients(symptom1, disease) to try to determine which diseases symptom1 alone implies (with high confidence) AND
in some cases, the columns of a Table are really instances of another entity.
E.g., Images, e.g., Landsat Satellite Images, LS(R, G, B, NIR, MIR, TIR), with wavelength intervals in micrometers: B=(.45, .52], G=(.52, .60], R=(.63, .69], NIR=(.76, .90], MIR=(2.08, 2.35], TIR=(10.4, 12.5]
Known recording instruments (that record the number of photons detected in a small space of time in any a given wavelength range) seem to be very limited in capability (eg, our eyes, CCD cameras...).
If we could record all consecutive bands (e.g., of radius .025 µm) then we would have a relationship between pixels (given by latitude-longitude) and wavelengths. Then we could do ARM on Imagery!!!!
Netflix
Intro to Association Rule Mining (ARM)
Given any relationship between entities, T (e.g., a set of Transactions an enterprise performs) andI (e.g., a set of Items which are acted upon by those transactions).
e.g., in Market Basket Research (MBR) the transactions, T, are the checkout transactions (a customer going thru checkout) and the items, I, are the Items available for purchase in that store.
The itemset, T(I), associated with (or related to) a particular transaction, T, is the subset of the items found in the shopping cart or market basket that the customer is bringing through check out at that time).
an Association Rule, AC, associates two disjoint subsets of I (called Itemsets). (A is called the antecedent, C is called the consequent)
The support [set] of itemset A, supp(A), is the set of of t's that are related to every aA,
e.g., if A={i1,i2} and C={i4} then supp(A)={t2, t4} (support ratio = {t2,t4}| / |{t1,t2,t3,t4,t5}| = 2/5 )
Note: | | means set size or count of elements in the set.
The support [ratio] of rule AC, supp(AC), is the support of {A C}=|{t2,t4}|/|{t1,t2,t3,t4,t5}|=2/5
The confidence of rule AC, conf(AC), is supp(AC) / supp(A) = (2/5) / (2/5) = 1
Data Miners typically want to find all STRONG RULES, AC, with supp(AC) ≥ minsupp and conf(AC) ≥ minconf
(minsupp, minconf are threshold levels)
Note that conf(AC) is also just the conditional probability of t being related to C, given that t is related to A).
e.g., Horizontal TransT T(I) Tablet1 i1
t2 i1, i2, i4
t3 i1, i3
t4 i1, i2, i4
t5 i3, i4
T I
A
t1
t2
t3
t4
t5
i1
i2
i3
i4 C
Its graph
APRIORI Association Rule Mining:
Given a Transaction-Item Relationship, the APRIORI algorithm for finding all Strong I-rules can be done:1. vertically processing of a Horizontal Transaction Table (HTT) or 2. horizontally processing of a Vertical Transaction Table (VTT).
In 1., a Horizontal Transaction Table (HTT) is processed through vertical scans to find all Frequent I-sets (I-sets with support minsupp, e.g., I-sets "frequently" found in transaction market baskets).
In 2. a Vertical Transaction Table (VTT) is processed thru horizontal operations to find all Frequent I-sets
Then each Frequent I-set found is analyzed to determine if it is the support set of a strong rule.
Finding all Frequent I-sets is the hard part. To do this efficiently, APRIORI Algorithm takes advantage of the "downward closure" property for Frequent I-sets: If an I-set is frequent, then all its subsets are also frequent.
E.g., in the Market Basket Example, If A is an I-subset of B and if all of B is in a given Transaction's basket, the certainly all of A is in that basket too. Therefore Supp(A) Supp(B) whenever AB.
First, APRIORI scans to determine all Frequent 1-item I-sets (contain 1 item; therefore called 1-Itemsets),next APRIORI uses downward closure to efficiently find candidates for Frequent 2-Itemsets,next APRIORI scans to determine which of those candidate 2-Itemsets is actually Frequent,next APRIORI uses downward closure to efficiently find candidates for Frequent 3-Itemsets,next APRIORI scans to determine which of those candidate 3-Itemsets is actually Frequent, ...Until there are no candidates remaining (on the next slide we walk through an example using both a HTT and a VTT)
Horizontal Transaction Table (HTT)
minsupp is set by the querier at 1/2 and minconf at 3/4(note minsupp and minconf can expressed as counts rather than
as ratios. If so, since there are 4 transactions, then as counts, minsupp=2 and minconf=3):
or a Vertical Transaction Table (VTT)
(downward closure property of "frequent")
Any subset of a frequent itemset is frequent.
APRIORI METHOD: Iteratively find the Frequent k-itemsets, k=1,2,...
Find all strong association rules supported by each frequent Itemset.
(Ck will denote candidate k-itemsets generated at each step.
Fk will denote frequent k-itemsets).
ARM The relationship between Transactions and Items can be expressed in a
2 3 3 1 31-Iset supports
2 3 3 3Frequent (supp 2)
Start by finding Frequent 1-ItemSets.
2 1s, 3 2s, 3 3s, 1 4, 3 5s
TID Items100 1 3 4200 2 3 5300 1 2 3 5400 2 5
TID 1 2 3 4 5
100 1 0 1 1 0
200 0 1 1 0 1
300 1 1 1 0 1
400 0 1 0 0 1
TID Items100 1 3 4200 2 3 5300 1 2 3 5400 2 5
HTTitemset sup.
{1} 2{2} 3{3} 3{4} 1{5} 3
Scan DScan D
C1
TID 1 2 3 4 5
100 1 0 1 1 0
200 0 1 1 0 1
300 1 1 1 0 1
400 0 1 0 0 1
itemset sup{1 2} 1{1 3} 2{1 5} 1{2 3} 2{2 5} 3{3 5} 2
C2
Scan DScan D
itemset{1 2}{1 3}{1 5}{2 3}{2 5}{3 5}
C2 F3 = L3
Scan DScan D itemset sup{2 3 5} 2
P1 2 //\\ 1010
P2 3 //\\
0111
P3 3 //\\ 1110
P4 1 //\\ 1000
P5 3 //\\ 0111
BuildPtrees:Scan DScan D
L1={1}{2}{3}{5}
P1^P2 1 //\\
0010
P1^P3 2 //\\
1010
P1^P5 1 //\\
0010
P2^P3 2 //\\
0110
P2^P5 3 //\\
0111
P3^P5 2 //\\
0110
L2={13}{23}{25}{35}
P1^P2^P3 1 //\\
0010
P1^P3 ^P5 1 //\\
0010
P2^P3 ^P5 2 //\\
0110
L3={235}
F1 = L1
itemset sup.{1} 2{2} 3{3} 3{5} 3
itemset sup{1 3} 2{2 3} 2{2 5} 3{3 5} 2
F2 = L2
{123} pruned since {12} not frequent{135} pruned since {15} not frequent
Example ARM using uncompressed Ptrees (note: I have placed the 1-count at the root of each Ptree)
C3
itemset{2 3 5}{1 2 3}{1,3,5}
ARM
L3
itemset sup{2 3 5} 2
L1
itemset sup.{1} 2{2} 3{3} 3{5} 3
itemset sup{1 3} 2{2 3} 2{2 5} 3{3 5} 2
L2
1-ItemSets don’t support Association Rules (They will have no antecedent or no consequent).
Are there any Strong Rules supported byFrequent=Large 2-ItemSets (at minconf=.75)?
{1,3} conf{1}{3} = supp{1,3}/supp{1} = 2/2 = 1 ≥ .75 STRONGconf{3}{1} = supp{1,3}/supp{3} = 2/3 = .67 < .75
{2,3} conf{2}{3} = supp{2,3}/supp{2} = 2/3 = .67 < .75 conf{3}{2} = supp{2,3}/supp{3} = 2/3 = .67 < .75
{2,5} conf{2}{5} = supp{2,5}/supp{2} = 3/3 = 1 ≥ .75 STRONG!conf{5}{2} = supp{2,5}/supp{5} = 3/3 = 1 ≥ .75 STRONG!
{3,5} conf{3}{5} = supp{3,5}/supp{3} = 2/3 = .67 < .75 conf{5}{3} = supp{3,5}/supp{5} = 2/3 = .67 < .75
Are there any Strong Rules supported byFrequent or Large 3-ItemSets?{2,3,5} conf{2,3}{5} = supp{2,3,5}/supp{2,3} = 2/2 = 1 ≥ .75 STRONG!
conf{2,5}{3} = supp{2,3,5}/supp{2,5} = 2/3 = .67 < .75
conf{3,5}{2} = supp{2,3,5}/supp{3,5} = 2/3 = .67 < .75
No subset antecedent can yield a strong rule either (i.e., no need to check conf{2}{3,5} or conf{5}{2,3} since both denominators will be at least as large and therefore, both confidences will be at least as low.
No need to check conf{3}{2,5} or conf{5}{2,3} DONE!
2-Itemsets do support ARs.
ARM
E.g., the $1,000,000 Netflix Contest was to develop a ratings prediction program that can beat the one Netflix currently uses (called Cinematch) by 10% in predicting what rating users gave to movies. I.e.,
predict rating(M,U) where (M,U) QUALIFYING(MovieID, UserID).
Netflix uses Cinematch to decide which movies a user will probably like next (based on all past rating history). All ratings are "5-star" ratings (5 is highest. 1 is lowest. Caution: 0 means “did not rate”).
Unfortunately rating=0 does not mean that the user "disliked" that movie, but that it wasn't rated at all. Most “ratings” are 0. Therefore, the ratings data sets are NOT vector spaces!
One can approach the Netflix contest problem as a data mining Classification/Prediction problem.
A "history of ratings given by users to movies“, TRAINING(MovieID, UserID, Rating, Date) is provided, with which to train your predictor, which will predict the ratings given to QUALIFYING movie-user pairs (Netflix knows the rating given to Qualifying pairs, but we don't.)
Since the TRAINING is very large, Netflix also provides a “smaller, but representative subset” of TRAINING,
PROBE(MovieID, UserID) (~2 orders of magnitude smaller than TRAINING).
Netflix gave 5 years to submit QUALIFYING predictions. That contest was won in the late summer of 2009, when the submission window was about 1/2 gone.
The Netflix Contest Problem is an example of the Collaborative Filtering Problem which is ubiquitous in the retail business world (How do you filter out what a customer will want to buy or rent next, based on similar customers?).
Collaborative Filtering is the prediction of likes and
dislikes (retail or rental) from the history of previous expressed purchase of rental satisfactions (filtering new likes
thru the historical filter of “collaborator” likes)
ARM in Netflix?Movie User RentsUID
CNAME
AGE
Rating
Date
MID
Mname
Date
ARM used for pattern mining in the Netflix data?
In general we look for rating patterns (RP) that have the property RP true ==> MIDr (movie, MID, is rated r).
Singleton Homogeneous Rating Patterns: RP = Nr
Multiple Homogeneous Rating Patterns: RP = N1 r & ... & Nk r
Multiple Heterogeneous Rating Patterns: RP = N1 r1 & ... & Nk rk
Related Documents
