ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding 1 Chapter One Introduction Chapter 1. Introduction - About the course - Motivations and examples - A brief history and course organization - Relation to other courses - Some basic concepts
18
Embed
Chapter 1. Introduction Chapter One Introductionise.tamu.edu/inen614/Chapter 1.pdf · Chapter One Introduction Chapter 1. Introduction ... 13, 1994) that contains eight ... - robust
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
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
1
Chapter One Introduction
Chapter 1. Introduction
- About the course
- Motivations and examples
- A brief history and course organization
- Relation to other courses
- Some basic concepts
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
2
Course syllabus
• ISEN 614 Advanced Quality Control
(Anomaly and Change Detection)
• Instructor: Dr. Yu Ding
• Course website: http://ise.tamu.edu/inen614
• Syllabus
• Course project
• Academic IntegrityAggie Honor Code: “An Aggie does not lie, cheat,
or steal or tolerate those who do.”
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
3
What is the course about?
• Fundamentally, this course is about concepts and
methodologies of change and anomaly detection.
• People usually try to detect “abnormal” from
“normal”.
• Examples of “abnormal” or anomalies: in quality
control, a bad product; in security applications,
criminals and terrorists; in healthcare application,
medical errors and disease outbreaks.
• In a process or environment, the detection of
existence of “abnormal” is anomaly detection; the
detection of the moment where “abnormal” appears
is change detection.
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
4
Example 1.1: A univariate process in quality control (QC)
• Example 1.1: Frozen orange juice concentrate is
packed in 6-oz cardboard cans. These cans are
formed on a machine by spinning them from
cardboard stock and attaching a metal bottom panel.
As part of the QC process, people need to inspect
possible leak either on the side seam or around the
bottom joint. There are 30 samples, each of which
has 50 cans were selected at half-hour intervals over
a three-shift period.
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
5
Example 1.1 (continued)
• Data table
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
6
Example 1.1 (continued)
• Data and chart
• Question: Is the fluctuation an unavoidable part of the underlying
process or is it an indication of some kind of process change?
0 5 10 15 20 25 30
4
6
8
10
12
14
16
18
20
22
24n
um
ber
of
no
nco
nfo
rmin
g
can
s in
each
sam
ple
sample index
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
7
Example 1.2 Monitoring medical errors
• Example 1.2: In a hospital, there are three major types of
medical errors to be detected and avoided. Every day, a
medical safety personnel will inspect a sample of 30 cases
and record the errors happened in each category.
• The person will send out an alert if s/he believes the error
rate increases out of the expected range and that it is likely
caused by a systematic root cause.
• However, an alert should be triggered only when it is
justifiable or “significant” in some sense. Otherwise, too
many alerts will practically shut down the hospital.
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
8
Example 1.2 (continued)
• Data table (multivariate)
• Question: when to send out alerts and which errors are the main
contributors?
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
9
Example 1.3 Scanning for unusual clusters
• Example 1.3: Over a five year period, 1991 to 1995, there
were 19 cases of a particular type of cancer reported in a
city. In reviewing the data, the epidemiologist notes that
there is a 1 year period (from April 4, 1993 through April
13, 1994) that contains eight cases.
• Question: Given 19 cases over 5 years, how unusual is it to
have a 1 year period containing as many as eight cases?
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
10
Example 1.4 Early warning system for West Nile virus
• Example 1.4: Since 1999 West Nile virus (WNV) outbreak in
New York City, which caused thousands of human infection
and 59 severe cases including 7 deaths, health officials have
been searching for an early warning system that could have
prevent human illness and death.
• In the summer of 2001, the New York City Department of
Health and Mental Hygiene established a citywide network of
adult mosquito traps, sentinel bird flocks, and system for
reporting, collecting, and testing dead birds.
• Health officials try to use the collected data and the pattern
embedded therein to set off public health alerts enough time
before onset of human cases.
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
11
Example 1.4 Early warning system for West Nile virus
• Question: Whether and when to
send out the alert? And where
is the potential outbreak epi-
center(s)?
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
12
Change and anomaly detection
• The objective of change and anomaly detection is to answer
the question: How “strange” is everything that has
happened in the last hours/days/months, given the historical
and recent observations?
• The answer to that question helps determine a proper
subsequent action, including, for example, active and more
intense data collections; stop a production process; alter the
public; dispatch interdiction force (police and other law
enforcements) etc.
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding
13
A brief history
• Two major branches of development:
- In the applications of public-health surveillance
- In the applications of industrial quality control
• Many tools are shared between these two areas but an
essential difference is that the public-health surveillance
applications deal primarily with discrete data, while the
industrial QC deal by and large with the continuous data.
• A recent new branch of relevant research is in computer
science because of the database applications and data-
mining demands.
- A Carnegie Mellon University group: Andrew Moore
and WSARE (What’s Strange About Recent Event)
ISEN 614 Advanced Quality Control (Anomaly and Change Detection) Spring 2008 Dr. Yu Ding