Application of Feature Engineering with Classification Techniques to Enhance Corporate Tax Default Detection Performance

摘要

The objective of this work is to propose a methodology that is helpful in analyzing tax data and predict significant features that cause tax defaulting. In this work, we gathered a Finnish tax default data of different firms and then split it according to primary and transformed feature sets. Different feature selection techniques were used to explore significant feature sets. After that, we applied various classification techniques into primary and transformed data sets and analyzed experimental outcomes. Besides, almost all classification techniques are represented the highest results for correlation-based feature selection subset evaluation, information gain feature selection and gain ratio attribute evaluation techniques. But, information gain feature selection is found as the most reliable feature selection method in this work. This analysis can be useful as a complementary tool to assess tax default factors in corporate sectors.