| Popis: |
A variety of problems must be overcome for a system that learns from examples to be useful. Such problems include reducing the dependency on the order of presented examples; reducing the number of examples required to learn a concept; pruning the generalization space; handling both conjunctive and disjunctive concept descriptions; anddealing with noisy training instances. This paper presents a system that effectively deals with many of these problems in a real-world domain by actively participating in the example selection process. 1. INTRODUCTION An active area of research in machine learning involves systems that induce concept descriptions from examples. Some well known inductive systems include Winston's program1 which learns structural descriptions from examples, Meta-DENDRAL2, a program that discovers cleavage rulers for mass spectroscopy, INDUCE3, a general method for learning disjunctive structural descriptions, and 1D34, a system that develops classificationrules by synthesizing decision trees from examples. Syntactic learning by induction from |
