Abstract—This paper presents the development of an efficient Back-propagation Artificial Neural Network (ANN) Algorithm suitable for prediction of occurrence of motor insurance claims. So, first an ANN Algorithm is created and then tested for standard engineering curves for efficiency and effectiveness. Then the motor insurance claims data based on past claim settlement records is sorted and a representative data is built. Using this data a predictive model is created by applying ANN Algorithm. Then the model is validated with the help of a new data set again representing the motor insurance claims sample set. It is concluded that the ANN prediction model developed here can be effective tool for prediction of the future motor insurance claims based on the pre-existing data. There is a scope for improvement of the Algorithm for faster convergence and higher accuracy. At the same time development of systematic procedures for selection of representative data is required.
Index Terms—Back propagation, motor insurance claims, predictive model, Steepest Descent Optimization.
F. Ashwini Bapat is a student of National Insurance Academy School of Management, Pune-45, India. (Contact No: +91 9226930122, e-mail id:firstname.lastname@example.org).
S. Dr. Prakash Bapat is Professor, Mechanical Engg. Department, Cummins College of Engineering for Women, Pune-52, India. (Contact No:+91 9273303942, e-mail id: email@example.com)
Cite: First Ashwini Bapat and Second Dr. Prakash Bapat, " Development and Testing of an Efficient Artificial Neural Network Algorithm and Its Effectiveness for Prediction of Insurance Claims," International Journal of Innovation, Management and Technology vol. 1, no. 3, pp. 286-291, 2010.