Multiple Imbalanced Attributes in Relational Data

Abstract

Real-world data is frequently saved in relational database systems with varying amounts of relevant properties. Unfortunately, most classification approaches are designed for learning from balanced non-relational data, and they are mostly used to categorise a single characteristic. We present a strategy for learning from relational data with the express objective of identifying several unbalanced characteristics in this work. We expand a relational modelling technique (PRMs-IM) intended for unbalanced relational learning to deal with multiple imbalanced attribute classification in our approach. We solve the problem of classifying multiple imbalanced attributes by incorporating the "Bagging" classification ensemble into the PRMs-IM. We demonstrate the effectiveness of our approach in predicting student performance using real-world imbalanced student relational data.