Multi criteria decision making framework for life insurance recommender system

Abstract

This research presents a Multi-Criteria Decision Making framework for providing unbiased and user-preference oriented Life Insurance Policy recommendations in a user friendly way. A life insurance recommender prototype is developed and validated by testing different phases of proposed life insurance recommender framework. In phase I, policy data model is created comprising of 165 life insurance policies of 05 life insurance companies of India. 07 policy constraints and 10 decision criteria of policies are identified. The policy data model consists of the performance matrix of all life insurance policies over the given set of policy criteria. Phase II provides a filtered list of life insurance policies in the form of clusters of different policy types. In phase III, the weight elicitation module of proposed HFUTO algorithm match the user s preferences with the policy criteria and weight matrix for the decision criteria is generated. In phase IV, the preference elicitation module of proposed HFUTO algorithm generate an overall performance score for the specified category of insurance policies. The proposed algorithm is tested and compared with other existing methods and gives a high degree of correlation. Sensitivity analysis performed also validates the proposed algorithm as quite sensitive to the preference change and is found to be consistent. In phase V, an interactive GUI based life insurance recommender desktop application (Insurance_4_YOU) is developed. The interface provides a login based access to the application and is validated by domain experts gaining accuracy of 97.4%. It is further cross validated by 400 diverse users and 98.75% of users have assessed the life insurance recommendations as pertinent, aligning closely with their input preferences and priorities. Additionally, a substantial 97.5% of users have described the Insurance_4_YOU application as effortless and user-friendly. newline