Evaluating the Creditworthiness of a Client in the Insurance Industry Using Adaptive Neuro-Fuzzy Inference System

Abstract

The article deals with the issue of a client´s creditworthiness assessment in the insurance industry. The article aims to identify new factors related to a client´s creditworthiness, and to create an assessment model. The factors which have relations to a client´s creditworthiness were identified in the first research stage. These factors represent the inputs into the model. The assessment model of the client´s creditworthiness was created in the second stage. In the third stage, the model was verified and implemented. The neuro-fuzzy method was used for creation, verification and implementation of the model. Five variables were selected as the inputs including damages, insurance length, insurance penetration, annual earnings and 2nd degree liquidity. These input variables were divided into two categories based on their nature (insurance indicators, accounting indicators). Research results show that the proposed model was verified above input data and can be used as a tool for supporting decisions concerning a client’s creditworthiness in the insurance industry. The main contribution of the paper is the identification of new factors which have relation to a client’s creditworthiness and the creation of the assessment model which works with these new factors transferred to fuzzy variables. The proposed model differs from the current approaches primarily thanks to its complex, systematic and hierarchical ability to evaluate the newly identified factors related to a client’s creditworthiness as fuzzy variables. Thanks to the model, it is possible to automate and accelerate the process of evaluation of a client’s creditworthiness in the insurance industry. The knowledge gained from the evaluation model is immediately possible to use in the strategic management of insurance companies e.g. in marketing activities.The article deals with the issue of a client´s creditworthiness assessment in the insurance industry. The article aims to identify new factors related to a client´s creditworthiness, and to create an assessment model. The factors which have relations to a client´s creditworthiness were identified in the first research stage. These factors represent the inputs into the model. The assessment model of the client´s creditworthiness was created in the second stage. In the third stage, the model was verified and implemented. The neuro-fuzzy method was used for creation, verification and implementation of the model. Five variables were selected as the inputs including damages, insurance length, insurance penetration, annual earnings and 2nd degree liquidity. These input variables were divided into two categories based on their nature (insurance indicators, accounting indicators). Research results show that the proposed model was verified above input data and can be used as a tool for supporting decisions concerning a client’s creditworthiness in the insurance industry. The main contribution of the paper is the identification of new factors which have relation to a client’s creditworthiness and the creation of the assessment model which works with these new factors transferred to fuzzy variables. The proposed model differs from the current approaches primarily thanks to its complex, systematic and hierarchical ability to evaluate the newly identified factors related to a client’s creditworthiness as fuzzy variables. Thanks to the model, it is possible to automate and accelerate the process of evaluation of a client’s creditworthiness in the insurance industry. The knowledge gained from the evaluation model is immediately possible to use in the strategic management of insurance companies e.g. in marketing activities.