Application of ILP to cardiac arrhythmia characterization for chronicle recognition

Abstract

We propose to use ILP techniques to learn sets of temporally constrained events called chronicles that a monitoring tool will use to detect pathological situations. ICL, a system providing a declarative bias language, was used for the experiments on learning cardiac arrhythmias. We show how to obtain properties, such as compactness, robustness or readability, by varying the learning bias.