To increase revenue, insurance companies have to explore extra operating income sources beyond policy proceeds. Interest income from policy loans is currently growing, and is achieving significant attention. Policy loan seems a good option to extend insurance company earnings. However, unlike a regular financial institution, an insurance company is rarely a primary candidate of applicant for loan applications. Hence, insurance companies must strive voluntarily and aggressively for the attention of applicants. Understanding the characteristics of loan applicants would provide helpful information, and is the goal of this study.

This study proposes a mining model to enable insurance company to predict potential loan applicants. The proposed model is composed of two components, a business rule generator and a recommendation mechanism. The browser logs of online users are also analyzed, and the loan-related information dissemination is discussed. This study cooperates with an insurance company in Taiwan which suffers from the above problem. As the illustration scenario, the proposed model enables insurance company predictions and then contacts the potential loan applicant in advance. The application rate of policy loans is expected to rise, implying that the interest revenue for insurance companies would increase.