Classification on Participants Renewal Process in Insurance Company: Case Study PT XYZ

Abstract

Insurance is a form of risk management and one of the fastest-growing businesses. PT XYZ is a company that focuses on health and life insurance. One excellent product owned by PT XYZ is Managed Care (MC) Insurance and it dominates 64.5% of the company's premium income. However, MC has a high claim ratio value. Proven by there were 363 companies that have a claim ratio of more than 76% in 2020. The increase in the total claim ratio is due to the company has not been able to predict the claim ratio when the renewal company applies for an insurance participant. This study focuses on classifying participants on insurance renewal process so that company can be more selective to approve the participants. Participant selection can help a company to reduce the claim ratio. The proposed method is doing classification on insurance participants’ data using 3803 datasets with four attributes and five algorithms and find significant features when generating the model. The models will be validated using k-folds cross-validation with k=10, evaluation results show the accuracy of each algorithm as following, Naïve Bayes 70.00%, Support Vector Machines 67.00%, Decision Tree 95.40%, Logistic Regression 90.20%, and Neural Networks 79.30%. The results of the study recommend the Decision Tree algorithm with an accuracy of 95.40% for the classification of renewal companies that will join as insurance participants because it has a better accuracy value than other algorithms. Decision Tree shows that the most significant features in defining prospective company assessment is the average age.