Hypothesis Testing with Classifier Systems

This thesis presents a new ML algorithm, HCS, taking inspiration from Learning Classifier Systems, Decision Trees and Statistical Hypothesis Testing, aimed at providing clearly understandable models of medical datasets. Analysis of medical datasets has some specific requirements not always fulfilled by standard Machine Learning methods. In particular, heterogeneous and missing data must be tolerated, the results should be easily interpretable. Moreover, often the combination of two or more attributes leads to non-linear effects not detectable for each attribute on its own. Although it has been designed specifically for medical datasets, HCS can be applied to a broad range of data types, making it suitable for many domains. We describe the details of the algorithm, and test its effectiveness on five real-world datasets.