Naïve classification approach for insurance fraud prediction

Abstract

An approach which can be used for the prediction of future potentials on the basis of present information is known as prediction analysis. This study is relied on the fraudulent discovery in the insurance business. A number of approaches have been projected up to now for the fraudulent discovery in insurance sector. These approaches mainly rely on machine learning algorithms. The insurance fraud detection is the major issue of prediction analysis. The insurance fraud detection has three phases which are pre-processing, feature extraction and classification. The naïve bayes classification approach is proposed in this work for the insurance fraud prediction. The proposed algorithm is implemented in python and results are analyzed in terms of accuracy, execution time. Keywords: Insurance Fraud detection, voting method, naïve bayes.