Next:
Contents
Contents
List of Figures
List of Tables
Introduction
The problem
Contributions
Overview of this document
Personalized Information Filtering
Previous Work
Information Retrieval and Filtering
Information Retrieval
Information Filtering
Learning and Adaptation in IR and IF
Software Agents
Learning Agents for Information Filtering
The proposed approach
Comparison with previous approaches
The Algorithm
Representation
Document
Profile
Filtering Documents
Extracting Document Representations
Scoring Documents
Selecting Documents
Learning from Feedback
Feedback for retrieved documents
Programming by demonstration
Genetic Algorithm
Crossover
Mutation
New Generation
Newt: An Implementation
Introduction
The Graphical User Interface
The Main Window
Reading News Retrieved By The Agent
Providing Feedback for Articles Retrieved
Manual Browsing and Programming By Demonstration
Adding New Agents and Training
Population of Profiles
Displaying Profiles
Archiving Agents
The Learning Module
User Feedback
Genetic Algorithm
The Information Filtering Module
Extracting Document Representations
Assigning Feedback
Scoring and Selecting Documents
Efficiency Issues
Experimental Results
Tests with real users
Results
Questionnaires
Tests with simulated users
Specializing to User Interests
Scenario 1
Experiment 1
Results
Scenario 2
Experiment 2
Results
Adapting to Dynamic User Interests
Scenario
Experiment
Results
Exploring Newer Domains
Scenario
Experiment
Results
Testing the complete system
Scenario
Experiment
Results
Conclusions
Future Work
The Filtering Engine
Genetic Algorithm
Agent model
The User Interface
References
About this document ...