What is Data Mining

1
What is Data Mining?

• Presented by
• Shane Brown
• Erin Bader
• Susan Shanlever

Chapter 1 MIS 6473 Dr. Segall January 26, 2004
2
Outline of Topics - Chapter 1

• Data Mining Defined
• Using Data Mining to Solve Specific Problems
• What Data Mining Is Not
• Avoiding the Oversell
• Practical Advice before You Begin

3
Some General Facts

• Organizations
• Data Cannot Be Analyzed
• Information is undervalued or underutilized
• Large Volume in DB
• Benefits of Organization
• New Patterns or Trends

4
Success in the Application of Data Mining

• Four Examples
• Improved Marketing Campaigns
• Improved Operational Procedures
• Identifying Fraud
• Examining Medical Records

5
(No Transcript)
6
Examining Medical Records

• Important to the military
• 80 not from battlefield
• Used to predict medications needed
• Because of data mining patterns were discovered
• Chickenpox in ages 17-19
• Cancer from same recruiting location

7
What Data Mining Is NOT!

• The focus of the data mining process is to
  discover hidden patterns and trends.
• Once a pattern is identified, it can be described
  as a known quantity.
• Once it is discovered, the data mining process is
  finished.
• Analytic approaches that search data sets on the
  basis of known patterns are not doing data mining

8
What Data Mining is NOT!

• We do not regard techniques that require
  implementation of rules, predefined training
  examples, or automated supervised learning to be
  data mining approaches.
• The techniques are still useful, but they are not
  a part of data mining.

9
Analysis Versus Monitoring

• The majority of data mining applications are
  focused on analyzing information that has been
  previously collected.
• Data are static and represent the state of the
  world in some past interval of time.
• You can review the information at your own pace,
  confirming the accuracy of the data, making
  considered decisions about which patterns are
  important.

10
Analysis Versus Monitoring

• The data does not change while the analysis is
  being performed, therefore it is reliable and
  consistent.
• Time involved in the decision making process is
  not an issue.

11
Analysis Versus Monitoring

• Monitoring often involves online pattern matching
  operations in which incoming data are compared
  against a set of conditions or boundaries.
• Monitoring often occurs in real time and involves
  the processing of data that are continually being
  updated.

12
Analysis Versus Monitoring

• Predictive models and forecasters can be used to
  help identify critical values, unusual behaviors,
  and criteria data.
• These systems are not usually performing data
  mining since they are not discovering new
  patterns or classifications.
• True data mining is difficult, but not impossible
  in these types of environments.

13
Monitoring Credit Card Transactions

• Credit card companies have elaborate systems to
  curb the misuse of their services and identify
  purchases that do not fit the clients profile.
• Companies must distinguish between good and bad
  transactions.
• There are predefined patterns for bad
  transactions
• Gasoline purchases in a series

14
Monitoring Medical Billing Fraud

• CPT unbundling
• Each medical procedure has a 5 digit code
  associated with it
• Problem occurs when doctors submit the claim and
  break up one actual procedure into several
  smaller procedures, therefore charging more.
• Constitutes insurance fraud/ happens often

15
Marketing with Coupons

• Companies compile lists of items in a grocery
  store that consumers are likely to purchase both
  if they purchase one.
• When you check out and a coupon is printed, it
  usually matches something that was in your cart.
• This also relates to the placement of the items
  within the store (next to each other)

16
Avoiding the Oversell

• Data mining services by the year 2000 will reach
  20 billion.
• Data mining is interactive discovery
• Data mining is unique and challenging
• It is not a silver bullet solution for all your
  questions
• The approach must be constantly refined.

17
Practical Advice Before You Begin

• The field of data mining shows exceptional
  promise in terms of its potential contributions
  to a host of analytical applications.
• Susan will offer some cautionary words of advice
  on some real-world issues that can limit the
  utility of data mining engagements unless
  addressed directly.

18
Practical Advice Before You Begin

• Justifying the Data Mining Investment
• Expand Marketing Campaigns
• Reduce Fraud
• Improve Profits
• Based on our experiences, companies usually look
  for the investment made in data mining to be
  about 15-20 percent of value of estimated losses
  or expected improvements. p. 17

19
Practical Advice Before You Begin

• Working Efficiently Timeliness Is a Virtue
• Results in Days or Weeks
• A Barrier to Quick Results Lack of Access to
  Data Sets
• If there are no interesting patterns found
• Poor Selection of the Data Extracted or Analysis
• Poor Quality Control in the Original Collection
  Process

20
Practical Advice Before You Begin

• Establishing the Limitations of Your Data
  Resources
• Access Available Data Sources
• Accurate
• Well-Coded
• Properly Maintained
• Data does Not Need to be Online or Interactive
• Attain Permission to the Data

21
Practical Advice Before You Begin

• Defining the Problem Up Front
• Find What is Of Interest or Importance
• Do the Analysis in Stages
• Avoid Promising Too Much

22
Practical Advice Before You Begin

• Knowing Your Target Audience
• Keep Your Target Audience in Mind
• Degree of Detail will Change for Different Target
  Audiences

23
Practical Advice Before You Begin

• Anticipating and Overcoming Institutional Inertia
• Understand it may Difficult for an Organization
  to Act on the Results of Data Mining Analysis
• in making the decision to use data mining
  you should give consideration to the types of
  data available for analysis and the types of
  outcomes that will be most useful within the
  context of the particular application area. p. 24

24
Questions?
25
Shanes Question

• What are some example areas in applying data
  mining successfully? (p.7-12 WB)

26
Erins Question

• What are the major difference between analysis
  versus monitoring?

27
Susans Question

• When establishing limitations for data resources
  What are the two most important things to know?