Transcript Case Base Reasoning(CBR)

Integrated Case-Based and
Rule-Based reasoning
approaches for Insurance
Presented BY
Palwencha Nagraj Krishna
(04329801)
Guided By
Prof Rajendra M. Sonar
Content






Introduction to Expert System
Case Base Reasoning
Hybrid System
Different Integrated Approaches
Insurance Domain Approaches
iKen Research Project at IIT
Bombay
Problem Definition
Field Officer
Policy to be
given or not
Agent
Advice
Policy
Client
Expert System


Expert systems are defined as an intelligent computer
program that uses knowledge and inference
procedures to solve problems those are difficult
enough to require significant human expertise for
their solution.
Knowledge-based expert systems, or simply expert
systems, use human knowledge to solve problems
that normally would require human intelligence.
Dr. Peter R. Gillett, Associate Professor, Department of Accounting & Information Systems, Faculty of Management, Rutgers University
ES Components
Software
Inference
Engine
Knowledge
Engineer
Users
Working
Memory
Expert
Knowledge
Base
Spreadsheet
Database
Data
Hardwqre
James P. Ignizio, “Introduction to Expert Systems The Development and Implementation of Rule-Based Expert Systems”, McGraw-Hill International
Expert Systems Types

Rule-based Expert Systems

Frame-based Systems

Hybrid Systems

Model-based Systems
Knowledge Acquisition and
Validation

Knowledge Engineering (KE)




Knowledge Acquire
Knowledge Validate
Knowledge Represent
Inferencing
Knowledge Representation


A good knowledge representation
naturally represents the problem
domain
KR Type




OAV Triplet
Semantic Network
Frames
Rules
Production Rules


Production rules are normally IF-THEN
variety.
The alternate designation for IF-THEN
rules is that of condition-action
statements.
Inferencing with Rules: Forward and Backward
Chaining
Rule:
IF A (is true)
THEN B (is the case)
IF = Premise
THEN = Assertion (or conclusion)
Pattern Matching:
Is A true?
Has it even been set?
If not, how is it set?




Backward Chaining Goal Driven
Traveling by Tain from Mumbai to Solapur

What Train arrive in Solapur

Do any originate in Mumbai

If not, for each origination, what Trains end there?

And where do they originate ….
Forward Chaining Data driven
Travling by Train from Mumbai to Solapur

s leaving Denver – Destinations

Are any destinations Tokyo?

If not, from those non Solapur dests, what flights Trains?

Which of those go to Solapur?

……
Structure of
Expert Systems


Development Environment
Consultation Environment
Decision Support Systems and Intelligent Systems, Efraim Turban and Jay E. Aronson
6th ed, Copyright 2001, Prentice Hall, Upper Saddle River, NJ
Introduction to CBR




Case base reasoning
system use the
technique to match a situation or problem
description to a stored database.
Input is given by the user on the current
situation and the output is case retrieval to
the most similar match to the database.
The CBR engine first searches
for case
history that are similar to the given
description.
The main intention is to reuse previous
experiences for actual problems.
. Diagnostic Stratrgies, “Expert System Development Series Introduction to Case- Base Reasoning”, www.DiagnosticStrategies.com.
Case

A case has two parts:



A problem or a set of problems.
The solution of this problem
A question can influence one or more
case.
Technology used:

Case Indexing:
A CBR system ability to retrieve relevant
cases quickly and accurately from its case
base is its main power.
It build a structure that will return the most
appropriate case(s) at high speed.
Different methods(1)

Nearest neighbor:
The system would simply prefer cases that
match more features to a case that matched
fewer.
Different methods (2)

Induction:
Inductive approaches to indexing are useful
where the retrieval goal or case outcome is
well defined.
The output of the induction process is in
the form of a decision tress.
Different methods(3)

Knowledge guided:
A knowledge-guided approach uses
human knowledge to the induction
process by manually identifying known
case features that are considered
important
The CBR cycle
Ian Watson & Farhi Marir (1994), “Case-Based Reasoning: A Review”, Cambridge University Press, 1994. The Knowledge Engineering Review, Vol.
Hybrid Intelligent Systems



Hybrid Intelligent System is a
combination of two techniques with
more strength and less weakness
Almost every conceivable problem has
been approached using some form of
hybrid system.
Hybrid systems are universally better
than conventional approaches.
Suran Goonatilake, Sukdev Khebbal, “Intellegent Hybrid Systems “, Goonatilake Khebbal Editor
Integrated expert systems
and case-based reasoning
Indexing
Problem
RBR
Indexed Case Library
CBR
Combination
Solution
Andrew R. Golding , Paul S. Rosenbloom, “Improving accuracy by combining rule-based and case-based reasoning”, ELSEVIER, Artificial Intelligence,
Case Study
If occupation( c) = student then attentive ; ‘Student’ rule
elseif sex(c) = M and age(c) < 30 then inattentive ; ‘Young driver’
rule
elseif age(c) 2 65 then inattentive ; ‘Old driver’ rule
else attentive ; ‘Default’ rule
If address2( c) = New York, NY
or address2(c) = Los Angeles, CA then endangered ; ‘Hostile traffic’
rule
else neutral ; ‘Normal traffic’ rule
If inattentive(c) and endangered(c) then high-risk ; ‘High’ rule
elseif inattentive(c) or endangered(c) then medium-risk ; ‘Medium’
rule
else low-risk ; ‘Low’ rule
Case Study cont…….
Integrated approaches
Sequential Inference Type:


For RBR and CBR, one is run firstly and another
Secondly.
Knowledge Conversion Type:


Convert cases into rules.
Host-Assistant Type:


RBR as host and CBR as assistant
Integrated-Reasoning Type:


Reason separately with CBR and RBR, then
compare the two results to make a decision
Zhi-Wei Ni, Shan-Lin Yang, Long-Shu Li, Rui-Yu Jia, “Integrated Case-Based Reasoning”, Volume 3, 2-5 Nov. 2003 Page(s):1845 - 1849
ES and CBR in Insurance

It consist of three key process




data entry,
data revision and
evaluation of data by the expert
RuleMaster provides ways to develop
the rules in the system.
Research Project – iKen

IIT has developed a Hybrid AI shell
environment named as iKen Core by Prof
Rajendra M. Sonar

This shell has four techniques




Rule Based Reasoning
Case Based Reasoning
Genetic Algorithm
Artificial Neural Network
Future Study




Studying the potential applications of
insurance domain where hybrid
approach can be use.
Devising methodology for knowledge
acquisition, presentation and retrieval.
Selection/customization of proper tool.
Developing prototype systems.
Thank You