基于语义关联和信息距离的个性化推荐方法研究  

Research on Personalized Recommendation Method Based on Semantic Association and Information Distance

作　　者：黎雪微 应时[3] 洪伟[4] Li Xuewei

机构地区：[1]武汉大学信息管理学院,湖北武汉430072 [2]文华学院经管学部,湖北武汉430074 [3]武汉大学计算机学院,湖北武汉430072 [4]武汉体育学院体育工程与信息技术学院,湖北武汉430079

出　　处：《情报理论与实践》2019年第11期142-149,共8页Information Studies:Theory & Application

基　　金：国家自然科学基金项目“基于SaaS软件运行日志分析的软件性能问题的在线识别和诊断方法”的研究成果,项目编号:61672392

摘　　要：[目的/意义]传统推荐方法仅考虑用户过去的兴趣偏好,忽略了用户兴趣偏好的漂移性问题,使得推荐结果过于专门化,不能给用户提供新颖的推荐项目。[方法/过程]文章提出了一种基于语义关联和信息距离的个性化推荐方法,该方法将项目的信息量融入到传统的语义关联相似度中,从而实现了用户兴趣偏好的有益迁移,使推荐得到有效扩展,改善了推荐专门化问题。[结果/结论]通过设计实验验证了信息距离能够对推荐结果产生较大影响,提出的方法可以给用户推荐其感兴趣并且更有价值的项目。随着新项目的不断加入,项目的信息量会动态变化,系统会不断调整推荐列表以适应用户需求。[局限]不足之处在于模拟仿真实验下样本量不足引起的可信度问题,后续的研究将利用爬虫工具收集大数据进行算法测试,验证方法在大样本环境下的有效性。[Purpose/significance]Traditional recommendation methods only consider users’ past interests and preferences,ignoring the drift of users’ interests and preferences,which makes the recommendation results too specialized to provide users with novel recommendation items.[Method/process]This paper proposes a personalized recommendation method based on semantic association and information distance,which integrates the information content of items into the traditional semantic association similarity,thus realizing the beneficial transfer of user interest preference,effectively extending the recommendation,and improving the recommendation specialization.[Result/conclusion]Through design experiments,it is verified that the information distance can have a great impact on the recommendation results,and the proposed method can recommend more valuable items of interest to users.With the continuous addition of new items,the information content of the items will change dynamically,and the system will constantly adjust the recommendation list to meet the needs of users.[Limitations]The deficiency of this paper lies in the credibility problem caused by the insufficient sample size in the simulation experiment.The following research will use the crawler tool to collect big data for algorithm test and verify the effectiveness of the method in the large sample environment.

关 键 词：个性化推荐 语义相似度 本体 用户兴趣模型 信息量 语义关联 信息距离 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]