基于细粒度观点挖掘的个性化推荐方法研究

发布时间：2018-06-02 22:44

本文选题：细粒度观点挖掘 + 个性化推荐　；参考：《浙江理工大学》2017年硕士论文

【摘要】：随着互联网的蓬勃发展,电子商务的迅速扩张,在网上购买产品已经成为现代人的一种生活方式。互联网购物平台提供了种类丰富的产品,用户为了能够从中选择符合自己需求的产品,需要花费大量的时间和精力。个性化推荐技术的出现为此提供了一种非常有效的解决方法。个性化推荐计算能够模仿销售人员帮助用户选择心仪的产品。目前的个性化推荐系统,一般通过对产品评分数据和用户的历史行为信息进行分析,而产品评论中存在的大量有价值的信息常常被忽略掉。一方面,用户给出的评分是对所购买产品的一个整体印象,只利用这些评分不能满足其他消费者对产品各个方面特征的不同评价信息获取的需求。另一方面,现有的利用评论文本的推荐研究,只是利用观点挖掘与情感分析给出消费者对产品整体性的褒贬性,并没有将消费者对产品各个方面特征的评价和所持有的观点与个性化推荐相来结合起来。基于以上原因,本文基于细粒度观点挖掘技术提出一种个性化的推荐方法。本文方法是通过对于用户评论文本的细粒度观点挖掘,获得产品各个方面特征下的情感强度和情感倾向,然后利用k-means聚类方法和D-S证据理论对细粒度观点挖掘的结果进行处理获得评分。最后,利用此评分对用户提供个性化推荐服务。基于以上方法给出的推荐产品可以被消费者信服,更加适合消费者个性化的需求。
[Abstract]:With the rapid development of the Internet and the rapid expansion of e-commerce, buying products online has become a way of life for modern people. Internet shopping platform provides a variety of products, users need to spend a lot of time and energy in order to choose products that meet their needs. The emergence of personalized recommendation technology provides a very effective solution. Personalized recommendation computing can mimic sales people to help users choose the desired product. The current personalized recommendation system generally analyzes the product score data and the user's historical behavior information, but a large number of valuable information in product reviews are often ignored. On the one hand, the rating given by the user is an overall impression of the product purchased, and the use of these ratings can not meet the needs of other consumers for different evaluation information of each aspect of the product. On the other hand, the existing research on the recommendation of the review text only uses the viewpoint mining and the emotional analysis to give the consumers' praise and demerit to the product integrity. It does not combine the consumer's evaluation and view of each aspect of the product with personalized recommendation. Based on the above reasons, this paper proposes a personalized recommendation method based on fine-grained viewpoint mining technology. The method of this paper is to mine the fine-grained view of the user comment text to obtain the emotional intensity and tendency of each aspect of the product. Then k-means clustering method and D-S evidence theory are used to process the results of fine-grained viewpoint mining to get the score. Finally, this score is used to provide personalized recommendation service to users. Based on the above method, the recommended products can be convinced by consumers and more suitable for consumers' individualized needs.
【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.3

【参考文献】