随着智能手机的普及和移动社交网络的兴起，以商户为中心寻找潜在客户已经成为移动环境中重要应用之一。本项目旨在研究移动环境下以商户为中心的潜在客户定向问题。主要研究内容包括：1）如何有效地基于位置、用户生成内容和社交信息离线生成客户兴趣偏好；2）如何高效地根据客户兴趣偏好在线完成以商户为中心的潜在客户匹配；3）如何合理地在隐私保护约束下确保客户定向服务的可用性。采用的技术路线包括：利用基于层次分类主题树来处理客户及其好友的生成内容，从而挖掘用户兴趣偏好；提出多维混合索引结构和移动对象安全区域的延迟更新技术，从而减小大规模数据场景下的客户定向计算代价；研究移动环境下不确定性数据管理的近似算法，从而在客户保护隐私的约束下确保定向服务的可用性。本项目适应当前移动环境下新型电子商务服务的发展需求，如能取得预期成果，具有广泛的应用前景和商业价值。

With the popularization of smart phones and boom of mobile social networking, merchant-oriented customer targeting has become one of the most important applications in mobile environment. This proposal aims to study merchant-oriented customer targeting in mobile environment. It includes: 1) how to efficiently detect user preference in an offline manner based on location, user generated content (UGC), social relationship, etc. 2) how to effectively find the matched customers for merchants in an online manner; 3) how to rationally guarantee the availability of user targeting with privacy protection constrains. In this proposal, we employ hierarchical taxonomy to process UGC from users as well as their friends, in order to detect user preference; we propose multi-dimensional index on location, and deferred update technique based on the safe region of mobile object, in order to minimize the cost; we study approximate algorithm on uncertain data management in mobile environment, in order to ensure the availability of targeting service with privacy protection constrains. This project keeps step with current requirement on the development of novel e-commerce in mobile environment, which has broad application prospects and business values if granted.