Research On Association Rule Mining Algorithm For Bank Credit Analysis

In recent years,with the development of society,the banking business is diversified and personal credit business is gradually rising,so the profit margin is very large.At present,most banks mainly rely on staff to evaluate their customers according to their own experience to decide whether to lend or not,which is very subjective and has a certain impact on the evaluation results.If we can scientifically analyze the creditworthiness of those people in bank transactions through data,then we can lend to those people so that increase bank earnings.Among many mining algorithms,Apriori algorithm is the most widely used.In recent years,it has been applied to data mining in all walks of life.The traditional Apriori algorithm has some shortcomings.Firstly,the algorithm needs to scan the transaction database several times,which increases the I/O overhead.At each scan,there are too many itemsets to scan.In addition,there are too many candidate itemsets generated in the process of self-linking,and the running time of the algorithm is long.Thirdly,Apriori algorithm is a serial algorithm,which can not meet the needs of large data mining.In this paper,we first analyze the improved AprioriTid algorithm,then analyze the shortcomings of the algorithm,and propose an improved Apriori_Cut algorithm for item reduction and sorting.The algorithm scans the database only once and reduces the data set traversal.It also reduces the number of candidate itemsets and avoids the generation of useless candidate itemsets.This paper first introduces the background and significance of the topic,and then introduces the theoretical knowledge related to this study.Then the principle of Apriori algorithm is introduced.Then,we analyze the traditional Apriori algorithm and the improved AprioriTid algorithm.Aiming at the shortcomings of the algorithm,we improve the algorithm and propose the Apriori_Cut algorithm.Then,Apriori algorithm,AprioriTid algorithm and Apriori_Cut algorithm are transplanted to Hadoop platform to adapt to large data mining.Subsequently,the three algorithms are compared by experiments,which proves that the Apriori_Cut algorithm is superior to the other two algorithms.Finally,we take the credit data of a foreign bank as the research topic to preprocess these data.Then we use Hadoop to process the data,and make a web application program to mine the bank credit data using Apriori_Cut algorithm.Then we analyze the mining results in detail,and summarize who has higher credit degree.Mining results can guide banks to make reasonable lending.