Learning as hill-climbing search Michael J. Pazzani Department of Information and Computer Science University of California, Irvine Irvine, CA 92717 (714) 824-5888 pazzani@ics.uci.edu http://www.ics.uci.edu/dir/faculty/AI/pazzani
Classification Learning
Version Space
Hill Climbing Search
Classification Problems
A Preview of Results
Parameter Tuning for Nynex Max: Minimizing Misclassification Costs
Hill Climbing for Parameter Tuning
Results
Issues in Hill Climbing Search
ID-2-of-3 (Murphy & Pazzani, 1991)
ID2-of-3 Results
First-Order Inductive Learner FOIL (Quinlan, 1990)
Learning a clause
FOIL Example
Literal Search Space
FOCL- Pazzani & Kibler(1992)
FOCL search strategy
Finding an informative operationalization
A Hill-climbing approach to Operationalization
Deleting literals of an operationalization
Foreign Trade Negotiations
Rules Learned by FOCL
CalendarPredicting Meeting Location
Calendar
PPT Slide
Theory Revision
A lexical bias for theory revision
Semantic Heterogeneity
CLARUS Results Student Loan
Bayesian Classifiers
Joining Attributes
Searching for Dependencies Backward Elimination and Joining
BSEJ Results
Related Search Methods
Summary
Email: pazzani@ics.uci.edu
Home Page: http://www.ics.uci.edu/~pazzani