基于Web挖掘的个性化视频推荐系统设计与实现
|
摘要
伴随着蓬勃发展的视频业务及内容,海量的视频信息常常导致用户选择困难,视频推荐技术随之诞生。传统的协同过滤算法存有推荐精度不高以及系统自身冷启动等问题。文章中设计并开发出一种基于Web挖掘的个性化视频推荐系统。该方法是采用用户Web日志模式进行分析,搜聚用户行为与属性建立效用数据矩阵,生成目标用户兴趣模型,在针对稀疏数据处理中采用PCA方式进行数据降维处理,也将内容和协同过滤的推荐优势同构化形成混合推荐算法,改进相似度计算方式。最终,实验结果验证了基于内容和协同过滤的混合推荐算法的平均绝对误差(MAE)比传统的基于内容或协同过滤算法整体降低了15%和6%。最后,利用Python Web技术和文章改进的算法搭建了电影推荐的原型系统。
With the booming video business and content,the massive video information often leads users to choose hardly,video recommendation technology is emerged. The traditional collaborative filtering algorithm has the problems of low recommendation accuracy and cold start-up of system.This paper designs and implements a video-recommending system that is based on Web mining. Using Web log to collect the users' behavior and attributes,namely establishing the users' interest model. In the model,using PCA to make dimension reduction in the sparse data processing and combined with the content and the integration of collaborative filtering recommendation algorithm,improving the methods of similarity calculation. The experimental results show that the mean absolute error(MAE)of the hybrid recommendation algorithm based on content and collaborative filtering is 15% and 6% lower than the traditional content-based or collaborative filtering algorithm. Finally,based on the technologies of Python Web,combining with theimprovement algorithm of this thesis,we build the prototype system of film recommendation.
With the booming video business and content,the massive video information often leads users to choose hardly,video recommendation technology is emerged. The traditional collaborative filtering algorithm has the problems of low recommendation accuracy and cold start-up of system.This paper designs and implements a video-recommending system that is based on Web mining. Using Web log to collect the users' behavior and attributes,namely establishing the users' interest model. In the model,using PCA to make dimension reduction in the sparse data processing and combined with the content and the integration of collaborative filtering recommendation algorithm,improving the methods of similarity calculation. The experimental results show that the mean absolute error(MAE)of the hybrid recommendation algorithm based on content and collaborative filtering is 15% and 6% lower than the traditional content-based or collaborative filtering algorithm. Finally,based on the technologies of Python Web,combining with theimprovement algorithm of this thesis,we build the prototype system of film recommendation.
引文
[1]倪霓,罗毅.关于视频推荐系统数据处理方法的分析[J].软件,2014(2):123-124.
[2]荣辉桂,火生旭,胡春华,等.基于用户相似度的协同过滤推荐算法[J].通信学报,2014(2):16-24.
[3]朱扬勇,孙婧.推荐系统研究进展[J].计算机科学与探索,2015,9(5):513-525.
[4]刘金勇.WEB数据挖掘在电子商务中的研究应用[J].网络安全技术与应用,2013(9):25-26.
[5]方传霞,闫仁武.基于Web挖掘的电子商务推荐系统研究[J].电子设计工程,2015(11):30-32.
[6]郑小坤,蔡杰,李书豪,等.基于用户行为日志的网站推荐[J].数字技术与应用,2017(6):231-232.
[7]程娟.基于云同步的视频推荐系统的系统实现[D].上海:上海交通大学,2015.
[8]李英壮,高拓,李先毅.基于云计算的视频推荐系统的设计[J].通信学报,2013(s2):138-140.
[9]黄莹,宋伟伟,邓春玲,等.协同过滤算法在电影推荐系统中的应用[J].软件导刊,2015,14(8):92-93.
[10]贾忠涛.基于协同过滤算法的电影个性化推荐系统设计与实现[J].软件导刊,2015(1):86-88.
[11]王越,程昌正.协同过滤算法在电影推荐中的应用[J].兵器装备工程学报,2014,35(5):86-88.
[12]徐佳.基于情境感知的个性化电影推荐[J].商,2016(3):203-204.
[13]钟前建,崔怀林,贾西平.基于Mahout框架的协同过滤视频推荐算法[J].广东技术师范学院学报,2016,37(8):14-19.
[14]胡于响.基于Spark的推荐系统的设计与实现[D].杭州:浙江大学,2015.
[15]唐真.基于hadoop的推荐系统设计与实现[D].成都:电子科技大学,2013.
[16]汪祥,贾焰,周斌,等.基于交互关系的微博用户标签预测[J].计算机工程与科学,2013,35(10):44-50.
[2]荣辉桂,火生旭,胡春华,等.基于用户相似度的协同过滤推荐算法[J].通信学报,2014(2):16-24.
[3]朱扬勇,孙婧.推荐系统研究进展[J].计算机科学与探索,2015,9(5):513-525.
[4]刘金勇.WEB数据挖掘在电子商务中的研究应用[J].网络安全技术与应用,2013(9):25-26.
[5]方传霞,闫仁武.基于Web挖掘的电子商务推荐系统研究[J].电子设计工程,2015(11):30-32.
[6]郑小坤,蔡杰,李书豪,等.基于用户行为日志的网站推荐[J].数字技术与应用,2017(6):231-232.
[7]程娟.基于云同步的视频推荐系统的系统实现[D].上海:上海交通大学,2015.
[8]李英壮,高拓,李先毅.基于云计算的视频推荐系统的设计[J].通信学报,2013(s2):138-140.
[9]黄莹,宋伟伟,邓春玲,等.协同过滤算法在电影推荐系统中的应用[J].软件导刊,2015,14(8):92-93.
[10]贾忠涛.基于协同过滤算法的电影个性化推荐系统设计与实现[J].软件导刊,2015(1):86-88.
[11]王越,程昌正.协同过滤算法在电影推荐中的应用[J].兵器装备工程学报,2014,35(5):86-88.
[12]徐佳.基于情境感知的个性化电影推荐[J].商,2016(3):203-204.
[13]钟前建,崔怀林,贾西平.基于Mahout框架的协同过滤视频推荐算法[J].广东技术师范学院学报,2016,37(8):14-19.
[14]胡于响.基于Spark的推荐系统的设计与实现[D].杭州:浙江大学,2015.
[15]唐真.基于hadoop的推荐系统设计与实现[D].成都:电子科技大学,2013.
[16]汪祥,贾焰,周斌,等.基于交互关系的微博用户标签预测[J].计算机工程与科学,2013,35(10):44-50.