摘要
对基于项目的传统的协同过滤算法进行了改进。传统的基于项目的协同过滤算法面临推荐效率低下和推荐精度不高的双重困难,为此,提出加权整合聚类分类预测方法,在数据处理和预测推荐过程中分别采用加权评分填充和重新定义相似性的办法提高推荐的准确度,并采用项目聚类的方法提高推荐效率,同时提出贡献度参数的概念对数据处理和预测推荐过程进行修正。采用MovieLens的数据集进行的实验对比,结果表明:改进算法能够明显提高协同过滤推荐算法的效率和精度,在数据比较稀疏的情况下依然能够保持较低的平均绝对偏差和较高的推荐效率。
This main objective of this study is to improve the item-based collaborative filtering algorithm.The traditional item-based collaborative filtering algorithm faced the double troubles on recommendation inefficiency and low accuracy.Therefore,the Weighted integration of clustering classification forecasting method is proposed in this paper.The method uses the re-weighted scoring approach and the re-definition of similarity approach in the data processing and in the process of forecasting recommendation respectively.It uses item clustering method to improve the recommendation efficiency.Meanwhile,the concept of contribution degree parameters to revise the data processing and the prediction of recommendation process is proposed.Through the comparison of the experiment with Movielens data sets,the results show that the improved algorithm can significantly improve the efficiency and accuracy of the collaborative filtering.,and also can maintain a low average absolute deviation and higher level of efficiency of the recommendation even when the data is relatively sparse.
出处
《重庆理工大学学报(自然科学)》
CAS
2010年第9期69-74,共6页
Journal of Chongqing University of Technology:Natural Science
关键词
加权整合
分类预测
贡献度参数
平均绝对偏差
推荐效率
weighted integration
classification and prediction
contribution degree parameter
average absolute deviation
recommendation efficiency