Learning Decision Rules by Randomized Iterative Local Search

Learning easily understandable decision rules from examples is one of the classic problems in machine learning. Most learning systems for this problem employ some variation of a greedy separate-and-conquer algorithm, which makes the rules order-dependent, and hence difficult to understand. In this paper, we describe a system called LERILS that learns highly accurate and comprehensible rules from examples using a randomized iterative local search. We compare its performance to C4.5, RIPPER, CN2, G-NET, Smog, and BruteDL, and show that it compares favorably in accuracy and simplicity of hypotheses in a number of domains.