基于Eclat算法的图书推荐系统仿真

Simulation of Book Recommender System Based on Eclat Algorithm

研究了图书馆的个性化推荐系统应用问题,针对常用的协同过滤技术不适用于大数据量的情况,在深入研究关联规则Eclat算法的基础上,为了高效挖掘和优先搜索有效信息,提出了一种改进算法,并将算法应用于图书推荐系统的仿真实验中,新算法充分利用了垂直数据表示和交叉计数的高效优势,直接在垂直数据表示的数据集上通过广度优先搜索和交叉计数产生频繁模式,通过对流通数据库中的借阅记录进行挖掘得到关联规则,产生读者感兴趣的书目。仿真结果表明算法可以在大数据量的情况下实现关联规则的高效挖掘,在图书推荐系统中取得了良好的应用效果。

In the research on library＇s personal recommender system,collaborative filtering is the common method in recommender system,but it cannot handle large data efficiently. An improved algorithm based on Eclat is given in the paper. The new algorithm is applied in simulation experiments of book recommendation system. The new algorithm makes use of a vertical data representation and cross-count high-performance advantage,generates frequent patterns directly in the vertical data representation of the data set through the breadth-first search and cross-count. Association rules are generated from library database. The simulation results show that the algorithm can achieve an efficient association rule mining on large data,and the knowledge generated by the new algorithm is effective for book recommendation.