Research on Rural Financial Data Mining and Intelligent Recommendation System Algorithm

A new rural finance product recommendation system based on mixed recommendation algorithm is proposed, which aims at improving the precision and practicability of recommendations. The system first uses a content-based recommendation method to screen out financial products that meet the basic needs of farmers based on their credit scores, income, loan history and other characteristics. Then, combined with the collaborative filtering algorithm; by analyzing the financial behavior of similar farmers, it recommends financial products selected by other farmers with similar characteristics. Finally, a weighted fusion strategy is used to combine the results of the two recommendation methods to obtain the best recommendation results. This study used 5, 000 farmers from several rural financial institutions to evaluate the performance of the traditional and mixed methods in precision, recall, F1. Compared with the traditional method, the proposed system improves the recommendation precision by 18% and the user's satisfaction by 15%. The advantages of the proposed algorithm in rural financial product recommendations are verified, and corresponding application suggestions are given.