Application Of Association Rule Mining In Bank Customer Credit Evaluation

With a new round of the wave of information technology, information technology became increasingly penetrated into all walks of life, many in the industry are slowly changing. The banking industry as a traditional industry, along with the information technology reform is also in constant development. At present, with the popularization of the Internet bank, more and more users can experience the convenient financial service. The life of people and the bank is more and more close, in order to give better customer service, and in order to improve the bank's risk management ability, the banking industry is to adapt to put forward higher requirements of bank customer credit evaluation system under the new background. Traditional risk management methods of statistics and related rules and regulations based on the application has not been enough to risk prevention and control to deal with the background of large data.This paper describes the technology of association rules in data mining related. According to the large number of user data accumulation of banking financial institutions, the use of data mining association rules mining technology for a large number of user data, can provide a reference for the user credit evaluation. In association rules mining, based on classical Apriori algorithm and its improved methods on the common, there is an optimization. Reference hash technology there are conflicts in the hash based process, set K entry in the hash to add counting, ensure that hash to each barrel of the K set is the only, and directly determine the frequent K itemsets, rather than by the candidate frequent K itemsets after a scan of the database to obtain frequent K itemsets, reduced one time of database scanning. Furthermore, in the first time of scanning database, the database information is stored in the form of an array, the future will no longer need to scan the database and can determined candidate itemsets support count directly by the array. The times of scanning database can be reduced, at the same time candidate set size can also be reduced. Finally in this paper, the classical association rules algorithm and the improved algorithm is applied to the bank customer credit data, the correctness of the improved algorithms is verified. With the efficiency analysis and comparison, it is concluded that the improved optimization algorithm in the efficiency has improved. The relationship between each index of customer credit data are analyzed, and obtains some useful customer credit data related knowledge, to facilitate the banking industry to build good customer credit evaluation according to the knowledge which is diged out from large amounts of data.