Genetic Adaptive Neural Networks for Prediction of Insurance Claims | |
---|---|
( Volume 1 Issue 6,December 2015 ) OPEN ACCESS | |
Author(s): | |
Fred Kitchens, Thomas Harris | |
Abstract: | |
In the insurance business, an underwriter’s two most important considerations are loss frequency and loss severity (probability of a loss and the financial value of the loss). Neural networks have been successfully applied to the insurance business in areas such as prediction of loss frequency, and prediction of bankruptcy. The objective of this study is to develop a neural network model to predict the severity of potential insurance losses in private passenger automobile insurance in the United States. The study predicts loss severity in private passenger automobile insurance using independent variables commonly available to the insurance underwriter based on the consumer’s application for insurance. A genetic adaptive neural network training algorithm is used to model the losses. Premiums are expected to be positively correlated with risk; therefore a linear model is developed as a benchmark, using the same data. These findings show that loss severity is more difficult to predict than frequency of loss. Improved results are likely to be found through alternative methodology, or data. |
|
Paper Statistics: | Total View : 471 | Downloads : 462 | Page No: 27-30 | |
Cite this Article: | Click here to get all Styles of Citation using DOI of the article. |