Bank Data Mining Based On Apriori Association Algorithm

The rapid development of the financial industry and the Internet has produced a large number of irregular financial data.Data analysis based on traditional finance and statistics is not enough to cope with such large-scale data,nor is it enough to customize benign financial policies.This paper discusses the algorithms related to association rules,studies the application of improved association algorithms in bank data mining,and effectively analyzes the potential association between irregular data for convenient application in practice.As Apriori algorithm needs to read the database many times and generate too many candidate frequent and complex itemsets,this paper proposes an association algorithm based on matrix optimization with reference to other improved association rule algorithms.The improved algorithm stores the data in a matrix,and uses 0 and 1to indicate whether the data exists in the database.The support count of items is obtained by calculating the number of 1 in the matrix vector,which simplifies the judgment of frequent item sets.After the new frequent itemsets are generated,they are transferred to the queue.By taking advantage of the first-in,first-out nature of the queue,the generated frequent itemsets are always orderly to avoid the repetition of itemsets.Compare the improved algorithm with the classic Apriori algorithm and FP＿growth algorithm verifies the advantages of the algorithm.For the generated frequent itemsets,this paper adopts a combined rule generation scheme and proposes a dynamic minimum confidence degree,which makes the association algorithm more close to the real data and effectively prevents the occurrence of too many useless rules and the omission of effective rules.After the rules are generated,the concept of lifting degree is added to ensure the rationality of the dynamic minimum confidence setting,so that the strong association rules finally provided are concise and effective,and can be applied in real life.In order to verify the effectiveness of the improved algorithm,this paper applies the improved algorithm and the improved rule generation method to the bank data,and carries out association analysis on the bank data.By comparing the mining time of frequent itemsets,it is verified that the improved algorithm improves the mining efficiency.At the same time,strong association rules are generated by a new rule generation method,which can effectively analyze bank data and provide strong theoretical support for banks and companies to formulate relevant financial policies.