Research And Application Of Association Rules In Data Mining

In data mining, association rules is an important research direction, and have achieved fruitful results in many ways. Association rules is to find the association or correlation among data items in the data, so as to help people to make decision. At present, the association rules are widely used in the field of marketing, financial market, Internet, transaction analysis etc..Studied in this paper, the main algorithms of association rules, association rules technology firstly expounds the background and the concept of association rules, and then introduces the association rules algorithm, and the Apriori algorithm in association rules and FP-growth algorithm to analyze the performance of Apriori algorithm, which has a significant historical position, with simple, easy to understand the characteristics of the low data requirements, but a large number of candidate itemsets and frequent scanning database transaction is two defects of the algorithm. The FP-growth algorithm by using the database transaction compression for FP-tree, then the recursive mining frequent itemsets, the algorithm scans the database only two times, and avoid the generation of candidate itemsets, but this algorithm consumes more memory.Aiming at the problems of Apriori algorithm, this paper presents the FP-Apriori algorithm, the algorithm will be transplanted to FP-tree Apriori algorithm, the algorithm and the Apriori algorithm, the use of frequent item set a priori knowledge of the nature, the iterative method of layer by layer search, exhausted data set of all frequent itemsets.The candidate item sets were selected, by visiting the FP-tree count at the same time, according to the FP-tree branch greatly, reduce access FP-tree through the pruning technology, so as to improve the efficiency of the algorithm, experimental results show the algorithm is feasible, and gives the analysis of the performance of the two algorithms.