HEALTH CLAIM INSURANCE PREDICTION USING SUPPORT VECTOR MACHINE WITH PARTICLE SWARM OPTIMIZATION

The number of claims plays an important role in the profit achievement of health insurance companies. Prediction of the number of claims could give significant implications for the profit margins generated by the health insurance company. Therefore, the prediction of claim submission by insurance users in that year needs to be done by insurance companies. Machine learning methods promise a great solution for claim prediction of health insurance users. There are several machine learning methods that can be used for claim prediction, such as the Naïve Bayes method, Decision Tree (DT), Artificial Neural Networks (ANN), and Support Vector Machine (SVM). The previous studies show that the SVM has some advantages over the other methods. However, the performance of the SVM is determined by some parameters. Parameter selection of SVM is normally done by trial and error so that the performance is less than optimal. Some optimization algorithms based heuristic optimization can be used to determine the best parameter values of SVM, for example, Particle Swarm Optimization (PSO) and Genetic Algorithm (GA). They are able to search the global optimum, easy to be implemented. The derivatives aren’t needed in its computation. Several researches show that PSO give better solutions if it is compared with GA. All particles in the PSO are able to find the solution near global optimal. For these reasons, this article proposes the health claim insurance prediction using SVM with PSO. The experimental results show that the SVM with PSO gives great performance in the health claim insurance prediction and it has been proven that the SVM with PSO gives better performance than the SVM standard.