Research On The Key Technology Of Information Association Based On Data Mining

With the popularity of the internet,more and more data are collected by large internet companies and research institutions,and the data contains a huge amount of valuable information that can be useful to researchers and business activists.The Application of data clustering analysis and association rules mining technology in data mining is very extensive,and these two technologies can be seen in business decision making,education industry and manufacturing industry.However,the traditional mining of association rules is relatively weak in mining knowledge of large datasets,so improving mining efficiency of association rules has become a popular topic in the industry.In order to improve the association rules mining technology,this paper introduced the data clustering analysis algorithm,and combined the clustering analysis algorithm with the traditional association rule mining algorithm to improve the mining efficiency.According to the technology of association rule mining and data clustering analysis,the following research was carried out.1.This paper analyzed the technical principle of association rule mining,introduced the main application fields of association rule mining and the basic steps of mining association rules,and gave the measurement method of association rules value.A method of mining association rules based on clustering analysis which first being applied in Web recommend was proposed.2.The feature task of clustering analysis and the method of performance measurement of clustering analysis algorithm were researched,analyzed the clustering principle of K-means algorithm based on centroid and density clustering DBSCAN algorithm,these two algorithm were applied to the apparel sales data,after compared the K-means algorithm and the DBSCAN algorithm,the advantages and disadvantages of the two algorithms were analyzed.3.This paper discussed the Apriori Algorithm of association rule mining and its improved algorithm,analyzed the mining efficiency of Apriori algorithm,Aprioritid algorithm and Apriorihybrid algorithm,and studied the FP-growth algorithm for improving frequent itemsets search,The construction principle of FP-tree is given,the Apriori algorithm and the FP-growth algorithm were applied to the mining of association rules of urban case data,the contrast experiment between the Apriori algorithm and the FP-growth algorithm was completed,the efficiency of FP-growth algorithm is higher than Apriori algorithm when the dataset is large.The parallelism of Apriori algorithm is higher than that of FP-growth algorithm.4.Finally,the K-means algorithm and the Apriori algorithm were combined,and the hybrid algorithm was applied to the Web page recommendation,which improved the efficiency and quality of the Web Recommendation.