Research of Improvement and Optimization on Association Rules Mining Algorithm
Xue Yong-sheng · Journal of Xiamen University · 2005
The most classical algorithm Apriori was studied and then separated into six steps:linking→cutting→generating C_k→scanning and counting→comparing→generating L_k. After that three improved measures were provided:1) judge if c∈C_k,2) scan while selecting,3) scan directly while bypassing cutting.Every measure has obvious advantages under special situation,and can effectively reduce the space occupied by candidate itemset and the time occupied by steps.Then a deep research was made on association rules generating process in detail,and a developed algorithm was given.The developed algorithm can decrease related computation quantity in large scale and improve the efficiency of the algorithm.