基于短期记忆与遗忘系数的用户个性化建模方法研究

发布时间：2018-02-22 17:38

本文关键词： 个性化搜索用户模型语义挖掘用户语境　出处：《计算机科学》2016年02期 　论文类型：期刊论文

【摘要】：搜索引擎的一个标准是不同的用户用相同的查询条件检索时,返回的结果相同。为解决准确性问题,个性化搜索引擎被提出,它可以根据用户的不同个性化特征提供不同的搜索结果。然而,现有的方法更注重用户的长时记忆和独立的用户日志文件,从而降低了个性化搜索的有效性。获取用户短时记忆模型来提供准确有效的用户偏好的个性化搜索方法被广泛采用。首先,根据基于查询关键词的相关概念生成短期记忆模型;接着,基于用户的时序有效点击数据生成用户个性化模型;最后,在用户会话中引入了遗忘因子来优化用户个性化模型。实验结果表明,所提出的方法可以较好地表达用户信息需求,较为准确地构建用户的个性化模型。
[Abstract]:One standard of search engines is that different users return the same results when they use the same query conditions. To solve the problem of accuracy, personalized search engines are proposed. It can provide different search results according to the individual characteristics of the user. However, the existing methods focus more on the user's long-term memory and independent user log files. In order to reduce the effectiveness of personalized search, the personalized search method which acquires user short-term memory model to provide accurate and effective user preference is widely used. Firstly, short-term memory model is generated according to the related concepts based on query keywords. Then, the user personalization model is generated based on the user's time series effective click data. Finally, the forgetting factor is introduced into the user session to optimize the user personalization model. The experimental results show that, The proposed method can better express the user's information requirements and build the user's personalized model more accurately.
【作者单位】：华东政法大学计算机科学与技术系;清华大学公共安全研究院;
【基金】：国家社会科学基金项目(06BFX051) 国家自然科学基金(6130202) 上海高校选拔培养优秀青年教师科研专项基金(hzf05046)资助
【分类号】：TP391.3

【相似文献】