Research And Application Of Data Mining Algorithm On The Marketing Of Bank’s Financial Product

How to apply the data warehouse and data mining technology to planning and selling in bank financial products is currently one of the most important fields in urgent need in China’s financial industry. This field includes the data mining technology research, more effective mining algorithm design, customer relationship management system to build.This paper does some research in key technologies during the implementation of the bank financing product sales analysis system. Also, an effective method of data mining in negative association rules is studied.Of many acts of bank customers, there is a positive, negative association rules. The traditional association rule algorithm only reflects the relationship between positive, without the negative correlation. This paper presents an effective algorithm (GA_PNAR), which is used to solve problem of the negative association of bank customer behaviour. The GA_PNAR algorithm firstly uses the Apriori algorithm to generate frequent item set, and then generates all negative association rules, through the NRGA algorithm based on correlation coefficient. When all the associations come out, genetic algorithm is adopted to optimize these rules. GA_PNAR algorithm is a very promising method to find optimal rules.The current methods for bank marketing mainly consider the basic attributes of the customer,not containing the customers’other attribute, such as value and behaviour. The designing of marketing systems also mainly concentrates on the customer segmentation, clustering analysis through the basic attributes of customers such as the investment horizon, risk preference. According to the results of cluster analysis, the most of systems combine the characteristics of clustering characteristics and financial products to customers, and provides financial solutions for customers. However, the information about customers, including the investment period, risk preference and other attributes are mainly gained through the test, which results in great inaccuracy and distortion. This paper presents a financial product correlation model based on the model of the customers. The model focuses on customer behavior rather than natural attributes. Compared to the conventional clustering analysis, customer analysis-financial products association rules, can provide a more accurate, specialized financial products Guidance for customers. In addition, in data preprocessing phase of the model uses the method of cloud model on the concept of stratification on numerical attributes, and the fuzz in the method can effectively solve the numerical hierarchical.For the domestic financial industry to implement the structured data mining technology, the deployment of enterprise data warehouse, improve customer classification, strengthening customer relationship management, market analysis, financial planning, dynamic market demand analysis and other aspects, the paper have a certain reference and guiding significance.