Using Data Mining to Save Time and Help You Look Like A Star

1
Using Data Mining to Save Time and Help You Look
Like A Star

• Irina Sered
• Executive VP
• www.wizsoft.com

2
What is Data Analysis

• Data analysis is the process of accessing the
  data looking for information to support a
  hypothesis.
• There are three types of tools
• Query tools
• Report writers
• OLAP tools

3
What is Data Mining

• Data mining is the process of revealing valuable
  information and discovering previously unknown
  relationships that exist in large amounts of
  data.
• The process complements standard statistical
  analysis to reveal patterns in data that
  translate into actionable information.
• Data mining sorts through the data and uncovers
  the critical information you need for strategic
  decision-making.

4
Data and Text Mining
WizRule
WizWhy
WizSame
5
What We Do

• WizSoft develops and markets software products
  based on mathematical algorithms for the business
  sector in the following fields
• data mining
• data auditing
• concept-based search engines
• ERP

6
Data Mining Tools for Auditors

• WizRule business rule detector
• WizSame duplicate records discovery
• WizWhy analyzer and predictor

7
Data Mining Tools for Auditors

• WizRule business rules detector for data
  auditing performs a complex analysis of the data
  revealing inconsistencies, errors and cases to be
  audited (investigated).

8
Data Mining Tools for Auditors

• WizSame duplicate records discovery, analyzes
  one data set, or compares records in two data
  sets and reveals similar records suspected as
  being duplicates.

9
Data Mining Tools for Auditors

• WizWhy analyzer and predictor, determines how
  the values of one field (dependent variable) are
  affected by the values of other fields.
• The resulting model consists of a set of
  IF-THEN, IF-THEN-NOT and IF-AND-ONLY-IF rules
  that describe valid associations between various
  values of the dependent variable and the values
  of other fields.

10
Data Mining Tools for Auditors

• WizWhy points out
• main patterns in the data
• summarizes the data
• lists unexpected rules
• lists unexpected cases in the data

11
Example of Unexpected Rule

• (basic rule 1) If BANK is Bank1
• Then DEFAULT is yes
• Rules probability 0.50
• (basic rule 2) If AGE is 17.00 20.00 (average
  18.61)
• Then DEFAULT is yes
• Rules probability 0.50
• (unexpected rule) If AGE is 17.00 20.00
  (average 18.61)
• and BANK is Bank1
• Then DEFAULT is yes
• Rules probability 0.05

12
Plain English Statements

• if there is a 50 chance that a loan from BANK1
  will default and
• if there is a 50 chance that a loan made to
  someone under the age of 20 will default,
• then it is unexpected that if a loan granted from
  BANK1 to someone under the age of 20 has only a
  5 chance of default.

13
In conclusion

• Auditors would want to investigate whether such
  unexpected rules are the result of spurious data,
  or whether they illuminate some underlying
  situation of which management was not aware.

14
Data Mining compliments Query Reporting

• Query based
• Requires a hypothesis
• Subjective interpretation
• Lots of reports and queries

• Data Mining
• No hypothesis necessary
• Objective analysis
• Critical few deviations

15
Summary Benefits

• WizRule, WizWhy, WizSame are complementary tools
  to any query-based audit
• Data integrity issues can waste auditors time
  learn what you dont know before relying on
  data
• Get an automated statistician to join the audit
  team
• Focus on the critical few deviations with no
  training and minimal effort
• Look like a star in your next audit by finding
  what no one else could find in the data

16
About Us

• Fully owned subsidiary of WizSoft Ltd. (1983)
• Headquarters in Tel-Aviv, Israel
• US office in Syosset, NY
• 30,000 customers world wide
• 170 employees
• 18 sales and services offices

• 516-393-5841
• www.wizsoft.com