Classification using small Training Sets based on Boolean Logic and Statistical Analysis

This work deals with the problem of producing a fast and accurate binary classification of data records, given a (possibly small) set of records that are already classified and have the same structure and nature of the former ones. The proposed approach is based on the framework of the Logical Analysis of Data (LAD), but enriched with information obtained from statistical considerations on the data. A number of discrete optimization problems must be solved in the different steps of the procedure. Accuracy of the proposed approach is compared to that of the standard LAD approach and of the Support Vector Machines approach (using the implementation LIBSVM, currently deemed one of the more effective) on publicly available datasets of the UCI repository. Encouraging results are obtained and discussed.