基于用户情境的高校图书馆书目协同过滤推荐研究

Research on Collaborative Filtering Bibliographic Recommendation of University Library Based on User Scenarios

高校图书馆进行书目推荐时忽略了用户的信息,导致面向读者的主动服务依然停留在探索的阶段。文章提出一种基于用户情境的书目协同过滤主动推荐策略。首先根据高校图书馆推荐系统用户的信息完整度将用户划分为三类,并针对不同用户类型进行了推荐流程分析;其次基于用户基本特征、历史行为及当前情境三个维度对用户情境进行建模,并利用贝叶斯网络推理构建基于用户情境的多维度偏好模型;最后通过应用实例和实验评分比较,验证基于多维偏好模型的书目推荐策略的有效性。

University libraries often ignore user's information in doing bibliographic recommendations,leaving the active reader-oriented service staying in the stage of exploration.The paper offers a bibliographic recommendation strategy that is collaborative filtering and user-scenario-tailored.Firstly,users are divided into three categories according to the degree of information integrity of the users in the recommendation system of university libraries,and the recommendation process is analyzed according to different types of users;secondly,user scenarios are modeled based on the three dimensions of users’basic characteristics,historical behaviors and current situation,and a multi-dimensional preference model based on user scenarios is constructed through Bayesian network;finally,the effectiveness and feasibility of the bibliography recommendation strategy based on the multi-dimensional preference model are verified by comparing experiments with applied cases.