摘要
针对当前地方志网站资源数量庞大,用户难以获取感兴趣的方志资源的问题,基于协同过滤技术,并结合Top N和改进的关联规则算法,提出一种混合推荐模型。该模型整合了Top N和改进的关联规则推荐以及协同过滤推荐的优点,利用方志标签对推荐结果进行筛选。实验结果表明,应用混合推荐模型不但能解决当前推荐技术普遍存在的用户评价信息稀疏、内容特征提取难度大、新用户推荐等问题,而且相比于单一的推荐技术在推荐质量上也有一定程度的提高。
In view of the current problem that there is such a large number of resources in local chronicles website where users are difficult to obtain interesting local resources,a hybrid recommendation model is proposed based on collaborative filtering technology,TopN and improved associated rules algorithm in this paper.The model integrates the advantages of TopN,improved associated rules and collaborative filtering recommendation,and selects recommended results using local chronicles labels.The experimental results show,that the application of hybrid recommendation model not only solves the problems of user evaluation information sparse,difficulty in content feature extraction,new user recommendations,and so on,but also improves the recommendation quality to a certain extent compared"with the single recommendation technology.
作者
黄涛
戴淑敏
成二丽
Huang Tao;Dai Shumin;Cheng Erli
出处
《国家图书馆学刊》
CSSCI
北大核心
2018年第2期14-19,共6页
Journal of The National Library of China
基金
本文系“十二五”国家科技支撑计划项目“地方志可视化技术研究与演示平台实现”(项目编号:2015BAK07B03)和“十二五”国家科技支撑计划项目“地方志资源调查与数字化加工规范研究”(项目编号:2015BAK07B01)的研究成果之一。
关键词
混合推荐技术
协同过滤
TopN算法
关联规则
Hybrid Recommendation
Collaborative Filtering
TopN Algorithm
Associated Rules