改进的Apriori算法在电子商务推荐系统中的应用

Application of Improved Apriori Algorithm in E-Commerce Recommendation System

针对目前电子商务的推荐系统不能适用于中小电子商务网站,文章使用改进的Apriori算法对电子商务交易事务数据库中的数据进行挖掘,首先对不同类的事务数据库中的最小支持度和最小置信度阈值进行区分设置,寻找最优值;然后对事务数据库中的数据进行稀疏性设置,转换成稀疏性矩阵的形式,以加快算法的执行效率,并每次都对与候选集中无关的项进行删除,再扫描修剪后的稀疏性矩阵,这样进一步提高挖掘效率。最后通过以某中小洁具用品电子商务网站的交易数据为对象,给出详细的操作方法和实验结果。

According to the recommendations of the electronic commerce system cannot apply to small and medium-sized e-commerce sites,this paper uses the improved Apriori algorithm to mine E-commerce trade affairs.Firstly,it sets the minimum support and the minimum threshold in all different kinds of affairs database,searches for the most optimal value,and then sets sparse solution and converts into a sparse matrix form to accelerate the algorithm efficiency,deletes the unrelated date in the candidate sets,further improves the mining efficiency.Finally it takes a small and medium-sized sanitary ware enterprise as the example,gives detailed operation method and experimental results.