BISC Decision Support System Evolutionary Computingbased MultiAggregator Fuzzy Decision Trees and op

1
BISC Decision Support System Evolutionary
Computingbased MultiAggregator Fuzzy Decision
Trees and optimization

• S. Souafi-Bensafi, G. Serag-Eldin, M. Nikravesh

BISC The Berkeley Initiative in Soft Computing
Electrical Engineering and Computer Sciences
Department
2
Outline

• BISC Decision Support System
• Objectives
• Structure
• Applications
• Extension to Web-Based BISC DSS
• Multi-Criteria Querying Model EC-based
  optimization

3
BISC Decision Support System

• Objectives
• Develop soft-computing-based techniques for
  decision analysis
• imprecise, uncertain and incomplete information
  and data
• consequence assessment in decision making and
  risk analysis
• human interaction, learning and adaptation
  through time

4
BISC DSS Components and Structure
Model and Data Visualization

• Model Management
• Query
• Aggregation
• Ranking
• Fitness Evaluation

Evolutionary Kernel Genetic Algorithm, Genetic
Programming, and DNA

• Selection
• Cross Over
• Mutation

Experts Knowledge
Input From Decision Makers
Model Representation Including Linguistic
Formulation
Data Management

• Functional Requirements
• Constraints
• Goals and Objectives
• Linguistic Variables Requirement

5
BISC DSS Process of Expert System
Knowledge Base
User
expertise is transferred and it is stored

• User Interface
• Dialog Function
• Knowledge Base Editor

Knowledge Refinement
users ask for advice or provide preferences
Expert Knowledge
Inference Engine
Data IF THEN Rule
inferences conclusion
advises the user and explains the logic
Recommendation, Advice, and Explanation
6
Applications

• Finance stock prices and characteristics,
  credit scoring, credit card ranking
• Military battlefield simulation and decision
  making
• Medicine diagnosis
• Marketing store and product display
• electronic shopping
• Internet provide knowledge and advice to large
  number of users
• Education university admissions
• Banking fraud detection

7
Extension to Web-based DSS

• Objective
• develop a generic framework for a multi-criteria
  fuzzy querying system
• Conceptual level
• Resembling natural human behavior - allowing
  approximation
• objects do not need to match exactly the decision
  criteria
• Implementation level
• Designed in a generic form to
• accommodate more diverse applications
• to be delivered as stand-alone software to
  academia and businesses.

8
Web-based DSS components
Application Template (AT)
UI
Fuzzy Search Engine (FSE)
Aggregators
Similarity measures
Norm-pairs
Evolutionary Computation
Membership functions
DB
9
Multi-Criteria Querying Model (1)

• Multi-Attribute Query

Scores
Database
Similarity calculation
Query
Query Answering Ranking based Selection
based (criteria number top answers) (criteria
threshold)
10
Multi-Criteria Querying Model (2)

• Aggregation-based model
• Data (x1, x2, , xk), Query (y1, y2, , yk)
• sj similarity(xj, yj), j 1, 2, , k
• Score Aggregation(s1, s2, , sk)

• parameters
• similarity measures
• norm-pairs
• aggregation operators
• weights
• aggregation model structure

Representation of user/expert preferences
11
First-order aggregation model

• Aggregation model simple aggregation operator
• Weighted aggregator user preferences attribute
  weighting

Optimization process find the optimal weights
Using GA.

• Model parameters learning using GA

12
Advanced multi-aggregation model

• Model description

• Parameters
• aggregators
• weights
• tree structure.

13
Model evaluation
Similarity calculation

• Fitness function combining
• distance D to maximize
• model structure size to minimize

14
Conclusions and perspectives

• Generic framework for multi-criteria flexible
  querying systems
• Generic EC-based optimization system
• Integrate data mining and knowledge discovery
  tools in the generic framework
• Develop more advanced evolutionary computing
  approaches for fuzzy modeling