DEVELOPMENT OF FUZZY EXPERT SYSTEMS FOR MODELING TACIT KNOWLEDGE IN COMMUNITY DECISION-MAKING

1
DEVELOPMENT OF FUZZY EXPERT SYSTEMS FOR
MODELING TACIT KNOWLEDGE IN COMMUNITYDECISION-MAK
ING

• D.S.Kalana Mendis, MPhil , CPhys, MIP (SL)Senior
  LecturerDepartment of Information
  Technology,Advanced Technological
  Institute,DehiwalaSri Lanka.
•  Prof. Asoka S. Karunananda, PhDFaculty of
  Information Technology, University of
  Moratuwa,Sri Lanka. 
• Dr. Udaya Samaratunga, MDSenior LecturerGampaha
  Wickramarrachi Ayurveda Institute,University of
  Kelaniya,Sri Lanka. 

2
Introduction

• Policy-makers and planners in communities have to
  make decisions in the face of varying degrees of
  uncertainty and risk
• Fuzzy logic and expert systems are two of the
  fundamental approaches that can be used to
  improve decision-making in relation to conditions
  of high levels of uncertainty
• Tacit knowledge is the key issue of knowledge
  modeling aspect
• The fundamental community problem is achieving
  purposeful, coordinated action from organizations
  comprising of many individuals.
• All knowledge is rooted in tacit knowledge
• Applying knowledge to decision making has a
  significant impact on community governance
• This paper presents a novel tool, where we have
  used multiple techniques to build an integrated
  expert system for developing an approach for
  modeling tacit knowledge

3
Good governance is

• A transparent decision-making process in which
  the leadership of an organization, in an
  effective and accountable way, directs resources
  and exercises power on the basis of shared
  values
• Marilyn Watt, A Handbook of NGO Governance
• A sharing of decision making so that power and
  resources do not accumulate in the hands of one
  person or a single group.

4
What is Tacit Knowledge?

• Difficult to formalize
• Practical
• Personal knowledge
• Context specific

5
usefulness of tacit knowledge

• The amount of tacit learning in an experience is
  the change the experience produces in an
  individuals ability to communities for their own
  decision making

6
RESEARCH PROBLEM

• The fundamental community problem is achieving
  purposeful, coordinated action from organizations
  comprising of many individuals because of tacit
  knowledge
• There are two dimensions to this problem first,
  the pure coordination problem second, the
  cooperation problem.

7
Challenges of using Tacit knowledge for effective
community decision making

• It is difficult discern and difficult to express.
  Examples include intuition, heuristics, and
  inherent talent.
• It enables people to have gut feelings that
  something is wrong or missing
• Lack of commitment of top management to sharing
  organizational tacit knowledge and absence of
  role models who are at the forefront of such
  knowledge sharing endeavors
• Lack of communication of vision and scope of a
  knowledge-sharing activity (Desouza, 2003)

8
Implications of not using Tacit knowledge for
effective community decision making

• Cannot represent individuals personal knowledge
  and experience
• Individuals viewpoints, beliefs
• Context specific skills and abilities
• Tacit knowledge in communities of practice tacit
  knowledge is personal knowledge

9
Tools for Community Design and Decision Making

• Computer Visualization Photo Montages
• Impact Analysis SCALDS
• Asset Mapping US Department of Urban Housing and
  Development EGIS
• GIS Analysis UGROW
• Predictive Modeling UrbanSim
• However tools for tacit knowledge in communities
  for community decision-making purposes has not
  been address in above-mentioned tools.

10
METHODOLOGY

• We postulate a fuzzy expert system as a tool for
  community governance through community decision
  making
• This emphasis to community decision making in a
  way by extracting, classification and reasoning
  tacit knowledge
• This enables community-decision makers to
  understand the structure of the target problem
  and identify its basic cause, which facilitates
  effective decision-making
• It is exploited the process of the new approach
  in following steps

11
EXTARCING TACIT KNOWLEDGE FROMCOMMUNITY

• The approach begins with by extracting tacit
  knowledge from community through domain expert
• Further modeling of tacit knowledge can be done
  as an interview between domain expert and the
  knowledge engineer
• Extracted tacit knowledge will be mapped in to a
  questionnaire

12
MODEL REFINEMENT FOR TACIT KNOWLEDGE MODELING

• Tacit knowledge is extracted from community to a
  questionnaire
• The questionnaire is analyzed using principal
  component analysis to reduce repetitiveness of
  questions in the questionnaire.

13
CLASSIFICATION OF TACIT KNOWLEDGE IN COMMUNITY

• Fuzzy logic is used to classify tacit knowledge
  in to various categories based on the
  questionnaire
• Likert scale is used to weight questions in the
  questionnaire

14
 REASONING IN COMMUNITY DECISION MAKING

• Fuzzy rules are generated for reasoning process
• Further reasoning for answers derived in tacit
  knowledge classification process
• Reasoning process for answers are given by the
  fuzzy rules

15
RESULTS

• Evaluated our approach using Ayurvedic medicine
  as a domain with tacit knowledge
• Classification of individuals through clinical
  examination in Ayurveda has been considered
• Classification of individual ?human constituents?
  is defined as a concept called ?prakurti
  pariksha.?
• Individual can be categorized into vata or pita
  or kapha based on the ?prakurti pariksha.?
• Method of analyzing constituents is not
  consistent

16
Extracting Tacit Knowledge in Ayurveda

• We mapped tacit knowledge regarding to analysis
  of constituents to a questionnaire
• It is consisted of 72 questions to analyse vata,
  pita and kapha

17
Removing Dependencies

• A pilot survey using a sample of 100 no. of
  laymen for statistical modeling
• Principal component analyzer has been used to
  remove dependencies
• It has been identified 25 principal components
  using SPSS

18
CLASSIFICATION OF HUMAN CONSTITUENTS

• Human constituents can be computed in to vata,
  pita and kapha in percentages
• Membership functions for vata, pita and kapha
  have been constructed using the out puts of
  principle component analyzer

19
(No Transcript)
20
Explanations For Derived Human Constituents

• Possible diseases can be occurred due to
  dominated constituent type
• Implemented through FLEX expert system shell

21
(No Transcript)
22
ANALYSIS OF RESULTS

• Tested with a group of 35 persons of
• Ayurvedic experts and Ayurvedic medical students
• The evaluation was conducted to see far the
  answers generated by the system matches with the
  identification by Ayurvedic experts and the
  students
• Systems ability to fine-tune the answers
• It is investigated that 77 of conclusions
  matches with the system and expert

23

• Derive constituents types in percentages while
  Ayurvedic experts obtain only the constituent
  type
• Find out possible diseases
• A self-assessment for finding constituents
• To find the effectiveness of minimum type in a
  diagnosis

24
CONCLUSION

• We suggest a tool for community decision-making
  using a fuzzy expert system
• It is with an emphasis to strategic concerns
• Contribute to community governance where
  communities make their own decision by
  extracting, classification and reasoning tacit
  knowledge
• Facilitates the linkage between knowledge
  management initiatives for policy inputs
• The system enables to achieve strategic goals and
  objectives of community governance