论文部分内容阅读
面对查询服务如何为用户提供满足需求的个性化推荐.提出一种基于自然语言进行评论分析、并提取特征属性进行多属性决策,为用户提供推荐排名策略,建立基于评论语义和Web挖掘技术的信息推荐系统实现个性化服务.解决了对同一商品的不同店铺之间的优劣比较和推荐,对各店铺的用户评论进行了主题抽取和情感分析,通过聚类成为“客户满意度”属性,与从店铺页面上爬取到的客观数据一起代入到推荐系统中进行计算.系统允许用户自主选择关心的属性及重要性排序,使得系统给出的推荐结果既能客观全面的反映店铺的状况,又能符合用户的评价偏好.
Faced with how query service can provide users with personalized recommendation to meet their needs, this paper presents a comment analysis based on natural language, extracts feature attributes for multi-attribute decision-making, provides users with recommendation ranking strategy, and establishes a semantic-based and Web mining technology based on comments Information recommendation system to achieve personalized service.To solve the advantages and disadvantages of different shops on the same product comparison and recommendation, the theme of each shop user comments and emotional analysis of the comments, through clustering become "customer satisfaction Attributes, along with the objective data crawled from the store page into the recommendation system for calculation.The system allows users to choose the attributes of concern and the importance of sorting, making the system gives the recommended results both objective and comprehensive reflection of the store’s Status, but also in line with the user’s evaluation preferences.