欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
个性化推荐技术是解决信息超载问题的最成功的技术之一.当前的个性化推荐系统中存在系统推荐质量不高、推荐算法可扩展性差、数据稀疏时推荐精度低等问题.针对这些问题,提出一种新的融合占有度和频繁度的协同过滤推荐算法.该算法利用占有度高斯权重来优化用户相似度和项目相似度,通过频繁度支持因子将基于用户的协同过滤算法和基于项目的协同过滤算法有策略地融合起来,实现目标预测评分的动态调节.在数据集movielens和netflix上的对比实验结果表明,该算法在目标邻居数目较少的情况下仍具有较高的推荐性能,相比其他参照算法的收敛速度更快,推荐精度较高,具有较好的扩展性.
英文摘要：
The personalized recommendation technology is one of the most successful techniques for solving the problem of information overload. The current personalized recommendation systems suffer from poor scalability,low quality and sparsity problems. In order to alleviate these problems,this paper proposes a novel frequency-based collaborative filtering algorithm fusing occupancy. The algorithm adjusts the similarity dynamically w ith the Gaussian w eight of occupancy. And then it integrates the user-based collaborative filtering algorithm w ith the item-based collaborative filtering algorithm strategically into an original algorithm w hich could dynamically adjust the target prediction score w ith the help of the frequency factor. Comparative experimental results on the w ell-know n film critic datasets movielens and netflix show that the algorithm in the case of sparsity performs better than other reference algorithms. It also converges faster and has higher scalability.
同期刊论文项目
同项目期刊论文
期刊信息
位置：