Evolution of Multi-adaptive Discretization Intervals for a Rule-Based Genetic Learning System

Genetic Based Machine Learning (GBML) systems traditionally have evolved rules that only deal with discrete attributes. Therefore, some discretization process is needed in order to teal with realvalued attributes. There are several methods to discretize real-valued attributes into a nite number of intervals, however none of them can eÆciently solve all the possible problems. The alternative of a high number of simple uniform-width intervals usually expands the size of the search space without a clear performance gain. This paper proposes a rule representation which uses adaptive discrete intervals that split or merge through the evolution process, nding the correct discretization intervals at the same time as the learning process is done.