Data Mining in Knowledge Management

1
Data Mining in Knowledge Management

• Fakulti Sains Komputer Teknologi Maklumat
• Fatimah Sidi
• 19/06/2002

2
Definitions of KM

• Address business problems particular to business
• creates and deliver innovative products or
  services
• managing and enhancing relationships with
  existing and and new customers, partners, and
  suppliers or
• administering and improving work practices and
  processes. (Tiwana, 2000)

3
Definitions of KM

• A system produces knowledge
• gathers information
• compares conceptual formulations describing and
  evaluating its experience, with its goals,
  objectives, expectations or past formulations of
  descriptions, or evaluations by comparison with
  reference to validation criteria (Firestone, 1998)

4
Definitions of KM

• A system maintains knowledge by continues to
  evaluate its knowledge base against new
  information by subjecting the knowledge base to
  continuous testing against its validation
  criteria.

5
Definitions of KM

• requires a knowledge base to begin operation
  where it enhances its own knowledge base with the
  passage of time because it is a self-correcting
  system, and subjects its knowledge base to
  testing against experience.

6
Definitions of KM

• re-badging of earlier information and data
  management methods
• Like any system of thgought that has value, both
  old and new and its combined new ideas with ideas
  that everyone has know all along (Prusak, 2001)

7
Definitions of KM

• Conclusion
• Knowledge Management is providing the growth of
  knowledge and also a new ways to channel raw data
  into meaningful information which in turn can
  become knowledge

8
Difference Between Data, Information Knowledge

• Data
• facts, numbers, or text
• operational or transactional data
• non operational data
• metadata - data about the data

9
Difference Between Data, Information Knowledge

• Information
• Collection of data is not information unless
  exist relation between the data
• Patterns, associations or relationships among
  data provide information

10
Difference Between Data, Information Knowledge

• Knowledge
• Information converted to knowledge about
  historical patterns and future trends
• Subset of information
• extracted, filtered or formatted in a very
  special way
• Subjected to and passed tests of validation

11
Difference Between Data, Information Knowledge

• Knowledge

Common sense knowledge is information that has
been validated by common sense experience
12
Difference Between Data, Information Knowledge

• Knowledge

Scientific knowledge is information
(hypotheses and theories) validated by rules and
tests applied to it by some scientific community
13
Difference Between Data, Information Knowledge

• Knowledge

Organizational knowledge is information
validated by rules and tests of the organization
seeking knowledge that improves organizational
performance
14
Difference Between Data, Information Knowledge

• Knowledge

leads to Wisdom arises when one understands the
foundational principles responsible for the
patterns representing knowledge.
15
Difference Between Data, Information Knowledge
( Gene Bellinger)
Context independece
wisdom
Understanding principles
knowledge
Understanding patterns
information
Understanding relations
understanding
data
16
Components KM technology framework (Tiwana, 2000)
Decision Support System
Workflow
Project Management
Data Mining
Knowledge Management
Document Management
Groupware
17
Components KM technology framework (Tiwana, 2000)

• Key Functions -
• Knowledge Flow
• Information mapping
• Information sources
• Information and knowledge exchange
• Intelligent agent and network mining
• Finding knowledge

18
Data mining in KM

• mechanism to appropriately cluster search results
  in different pre-specified content categories as
  specified in the knowledge map.
• Drill down into a relevant category without
  having to learn the subtleties of complex query
  languages and syntaxes

19
Definitions of DM

• Sometimes called data or knowledge discovery
• Process of analyzing data from different
  perspectives and summarizing it into useful
  information and
• Finding correlations or patterns among dozens of
  fields in large relational databases.

20
Definitions of DM

• (Holsheimer and Siebes, 1994)
• searching for relationships and global patterns
  that exist in large databases, but are hidden
  among the vast amounts of data.

21
Definitions of DM

• (Miller and Rohberg, 1996)
• tool that identifies and characterize
  interrelationships among multivariable dimensions
  without requiring a human to ask specific
  questions.
• looks for trends and patterns
• finds relationships and make prediction.

22
Definitions of DM

• (Han and Kamber, 2001)
• extracting or mining knowledge from large
  amounts of data.
• essential step in the process of knowledge
  discovery in databases, consists of an iterative
  sequence of the following steps
• Data cleaning

23
Definitions of DM

• Data integration
• Data selection
• Data transformation
• Data mining
• Pattern evaluation
• Knowledge presentation

24
How does DM work?

• Large scale information evolved transaction and
  analylitical systems separately
• DM provides link between the two
• Analyzes relationships and pattern in stored
  transaction data based on open queries.

25
How does DM work?

• Several types of analytical software available
• Statistical
• Machine learning and
• Neutral networks
• DM functionalities used to specify kind of
  pattern found in data mining task

26
Classification of DM

• Summarization (Holsheimer and Siebes, 1994)
• Association Rules
• Classification
• Clustering
• Prediction
• Sequential Patterns
• Similarity Search

27
Classification of DM

• Similarity Search (Algawal Swami, 1993)
• Outlier Anlysis (Han Kamber, 2001)
• Evolution Analysis

28
Major element in DM

• Extract, transform and load transactional data to
  DW
• Store and manage the data
• Provide data access to business analysts and
  information technology professionals
• Analyze the data
• Present the data

29
Levels of Analysis

• Artificial neural networks Non-linear predictive
  models
• Genetic algorithms
• Decision trees
• Nearest neighbor method
• Rule induction
• Data visualization

30
Objectives of the study

• To study the effective method of mining the
  knowledge in data mining
• To develop and implement the methods in mining
  the knowledge
• To test and measure its performance retrieving
  the knowledge

31
Thank You