Feature weighting by explaining case-based planning episodes

Abstract

We present a similarity criterion based on feature weighting. Feature weights are recomputed dynamically according to the performance of cases during planning episodes. We will also present a novel algorithm to analyze and explain the performance of the retrieved cases and to determine the features whose weights need to be recomputed. Experiments show that the integration of our similarity criterion in a feature weighting model and our analysis algorithm improves the adaptability of the retrieved cases over a period of multiple problem solving episodes.