项目名称: 基于领域知识和链路预测的个性化推荐研究
项目编号: No.71471169
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 管理科学
项目作者: 张玲玲
作者单位: 中国科学院大学
项目金额: 60万元
中文摘要: 大数据环境下,如何从信息的海洋中为用户提供个性化程度更高、更符合需求的商品和服务,已经成为电商在日益激烈的竞争环境下决胜的关键、商家和学者关注的焦点和热点研究方向。本研究针对现有个性化推荐中存在的重视推荐准确性忽视多样性、忽视领域知识、情境等问题,从知识管理、数据挖掘、营销管理、复杂网络交叉学科的角度,在研究个性化推荐机理的基础上,重点研究用户兴趣、行为、情境模型的建立,产品领域知识的获取、表示,产品-用户二图的建立方法,以及基于领域知识和链路预测的个性化推荐模型、算法和关键技术,并在此基础上研究对推荐结果进行二次或深层次挖掘,以更好地支持企业决策的技术和方法。该研究在理论上将领域知识、链路预测和个性化推荐三者结合,探索个性化推荐方法上的创新,从一个新的视角丰富了个性化推荐理论。同时,也可对电商企业提升个性化推荐的多样性、准确性和企业效益提供理论和实践的指导。
中文关键词: 个性化推荐;链路预测;领域知识;推荐多样性;可拓学
英文摘要: On the context of big date, how to provide the users with more personalized goods and service to meet their demand from the ocean of information is becoming the key to win out in an increasingly fierce E-business competition environment. Upon the researching basis of personalized recommend mechanism, this study will focus on the research of user interest, behavior and context modeling, product's domain knowledge obtain and representation, product-user bipartite constructing, and the personalized recommend model, algorithm and key technology innovation based on domain knowledge and link prediction, from the perspective of the inter-discipline of knowledge management, data mining, marketing management and complex network. Then we will do further research on recommend result's secondary mining and deep knowledge discovery to support enterprise decision. The research will theoretically combine the domain knowledge, link prediction and personalized recommend, and explore the method innovation on personalized recommend. Our research will enrich the personalized recommend theory from a completely new perspective and improve the diversity and accuracy of E-business personalized recommend, and guide the enterprise earnings improving theoretically and practically.
英文关键词: Personalized Recommend;Link Prediction;Domain Knowledge;Recommend Diversity;Extenics