位置相关的网络信息数据查询推荐模型与算法研究

The sales of Internet-worked mobile devices such as smartphones are skyrocketing. The mobile communication facilities are being improved. The smartphones are able to do variety of things. As a result of this development, the web is accessed increasingly by mobile users and mobile search has become a mainstream way of obtaining information. The mobile search market in China grows rapidly. Mobile search refers to that users search nearby relevant information via smartphones. However, it is difficult for users to accurately express their search intent via short keywords. Query suggestion in web search helps users to access relevant information without having to know how to precisely express their queries. Existing query suggestion techniques do not consider the locations of the users and the query results; i.e., the spatial proximity of a user to the retrieved results is not taken as a factor in the recommendation. However, the relevance of search results in many applications (e.g., location-based services) is known to be correlated with their spatial proximity to the query issuer. This project study the new problem of location-aware web query suggestion, the main research tasks include: (1) design a new suggestion model, which captures both the semantic relevance between queries and the spatial distance between the resulting documents and the user location, (2) new scheme for cold-start queries, and (3) to make the model scalable, design efficient algorithm for the computation of the suggested queries. The expected research results of this project will improve the mobile search technology in China and benefit the industry.

随着移动互联网设备如智能手机销量的迅速增长，移动通信设施建设步伐加快，手机技术的不断完善和功能的增加，用手机上网已成为一种获取信息的主流方式，中国移动搜索用户规模增长迅速。移动搜索是指用户通过智能手机查询用户当前位置附近的符合查询语义的空间网络信息数据。查询推荐作为一种帮助用户更好表达查询意图并提高查询结果质量的手段，在当前互联网搜索中得到广泛应用。但是现有的查询推荐方法忽略了用户的查询位置，无法保证推荐查询所获取的数据在用户的查询位置附近，因此在移动搜索中失效。本项目针对移动搜索中的推荐查询这一新问题，研究新的推荐模型和算法，使得推荐的查询能获取在空间上和语义上同时满足用户需求的数据。主要研究内容包括：（1）位置相关的查询推荐模型；（2）位置相关的查询推荐模型的抗冷启动机制；（3）大规模数据应用下，快速计算推荐查询的算法。本项目的预期研究成果将促进我国移动搜索领域的技术创新和产业应用。

当用户不确定如何精确表达信息检索需求时，搜索引擎中的关键字推荐技术可帮助用户找到相关信息。现有的关键字推荐技术没有考虑用户的当前地理位置，因此用户与检索结果的空间接近度不被视为推荐中的因素。然而，在许多已知应用（例如，基于位置的服务）中搜索结果的相关性与它们与用户的空间接近度相关。在这个项目中，我们设计了一个位置感知的关键字查询建议框架；提出了一个加权关键字 - 文档图，它表达了关键字查询之间的语义相关性以及结果文档和用户位置之间的空间距离；基于带重启的随机漫步模型，计算被推荐的查询与用户原查询之间的相似度，选择具有较高相似度的关键字查询作为建议。为了使所提出的框架可扩展，我们提出了一种基于分区的方法，它比基线算法在计算速度上快出一个数量级。我们在真实数据上评估了所提出方法的有效性和算法的性能。