摘要
推荐系统是目前在电子商务中用的较为广泛的一种技术。伴随着数据量的增大,评分矩阵的稀疏性成为了一大难题。对于评分数据较为稀疏的矩阵,提出了一种基于缺失值迭代预测填充的协同过滤算法。这种算法以迭代的方式对评分矩阵填充,直到缺失值个数恒定在某一数值。而在迭代的过程中,每一次用于填充计算的相似度度量又是依据均值填充后的相似度来动态计算的。说明该算法即可以降低数据稀疏性,又提高了用户相似度计算精度的问题。实验研究表明,利用该算法能够提高评分矩阵的密度,并降低了系统的推荐误差。
Recommendation system is a widely used technology in the electronic commerce.Along with the increase of the amount of data,sparsity of rating data become a big question.To improve sparsity of rating data more effectively,a collaborative filtering algorithm based on predicting and filling miss-data by interated is proposed.This method fills the rating data by iterated until the number of missing-data stably.During the iterating,the method of similarity analysis based on the result-data at last step.So not only this method improves sparsity of rating data more effectively,but else efficiently improves the accuracy of similarity analysis under the exreme sparsity of rating data.The experimental results show that this method can improve the quality of recommendation.
出处
《计算机与数字工程》
2016年第6期992-996,共5页
Computer & Digital Engineering
基金
国家自然科学基金(编号:11471161)资助
关键词
推荐系统
协同过滤
迭代
预测
相似度计算
缺失值填充
数据密度
recommendation
collaborative filtering
iteration
prediction
similarity computing
filling missing-data
data density