面向本地和外地用户情感分析推荐模型

Recommendation model of sentiment analysis for both home-town and out-of-town users

针对地理标签和评论信息的情感倾向对于推荐系统性能的影响,本文基于地理标签和用户评论情感分析提出有关兴趣点的推荐策略,并建立了一种基于内容的推荐模型.本系统首先对用户兴趣点信息进行有效的补充,并实现了用户兴趣点相似度度量.对无标签评论数据进行情感分析及挖掘,获取其情感倾向度.同时本系统结合了时间滑动窗口,更准确地把握用户评论和兴趣点的结合度.最终得到用户个性化推荐排名.本文方法涵盖了本地用户和外地用户的个性化推荐策略.通过实验数据表明,本文模型有效提高了推荐的准确度.

Aiming at the influence of the sentiment tendency of geographical position and comments on the performance of the recommendation system,this paper proposes a recommendation strategy for points-of-interest based on geographical position and sentiment analysis of user reviews,and establishes a content-based recommendation model.First,the system effectively supplements the user’s point-of-interest information,and realizes the similarity measurement of the user’s point-of-interest.Secondly,emotional analysis and mining of unlabeled comment data are carried out to obtain the sentimental tendency.At the same time,the system combines a time sliding window,and more accurately grasp the combination of user comments and points-of-interest.Finally,the personalized recommendation ranking of users is obtained.In this paper the method cover the personalized recommendation strategies of home-town users and out-of-town users.The experimental data shows that the model in this paper effectively improves the accuracy of the recommendation.