Insolvency Analysis On Electricity Billing Databases Using Bayesian Classifiers

Abstract

The present work verifies the applicability of Bayesian Classifiers over Databases from an energy distribution company. The purpose is finding patterns or profiles into determined energy consumption groups and to estimate the number of insolvent clients. The predictive computational system identifies patterns related to each client historic and projects probable behaviors. Insolvency predictions from a Bayesian Network Augmented Naïve-Bayes (BAN) Classifier are compared to results obtained by a Tree Augmented Naïve-Bayes (TAN) Classifier and a Naïve-Bayes (NB) Classifier, taking into account historical insolvency records. Validity is verified by comparing prediction errors. Conclusions suggest an adequate approach that offers arguments for elaborating effective commercial policies for reducing insolvency.

Fabrizzio Soares

Associate Professor and CS Chair

Fabrizzio Soares is a professor of Computer Science, Information Systems and Software Engineering at INF/UFG. His research interests include Computer Vision, Human Computer Interaction, Machine Learning and Programming topics. He is the leader of the Pixellab group, which develops solutions for accesibilty, Precision Agriculture, and Interactive Systems.