CAJ | 학술논문

图书馆提供信息化及智能化的信息咨询、推荐与提供服务是高校建设智慧化校园的重要组成部分.本文在充分分析高校数据孤立、业务流程低效、交互性差等实际问题的基础上,给出基于云计算和大数据等技术来构建统一、规范的图书个性化推荐与服务的统一平台. 将该系统划分为数据存储层、业务处理层和界面表示与交互层,并设计了每个层次上的系统功能. 进一步地,以业务处理层中的图书资料推荐算法为重点,设计了一种基于主题模型和关联规则挖掘的混合型推荐算法. 通过实验证实,该方法能比基于文本相似性的传统方法获得更高的推荐准确性.
도서관제공신식화급지능화적신식자순、추천여제공복무시고교건설지혜화교완적중요조성부분.본문재충분분석고교수거고립、업무류정저효、교호성차등실제문제적기출상,급출기우운계산화대수거등기술래구건통일、규범적도서개성화추천여복무적통일평태. 장해계통화분위수거존저층、업무처리층화계면표시여교호층,병설계료매개층차상적계통공능. 진일보지,이업무처리층중적도서자료추천산법위중점,설계료일충기우주제모형화관련규칙알굴적혼합형추천산법. 통과실험증실,해방법능비기우문본상사성적전통방법획득경고적추천준학성.
Libraries provide intelligent information advices and recommendations, which is an important part for the construction of wisdom campus. The current situation of campus construction is mainly reflected by the fol-lowing features: isolated data, low efficiency of business process and poor interaction, etc. Based on the deep anal-ysis, the paper proposes to build a unified and standardized platform for book service and personalized recommen-dation based on cloud computing and big data techniques. The whole platform is divided into three levels: data storage layer, business process layer and the interface representation and interaction layer. At the same time, the function in each layer is also designed. Furthermore, for the business of books and documents recommendation, the paper presents a hybrid algorithm by combining topic model and associate rule. Finally, the experiments con-firm that our method can achieve higher accuracy than conventional text similarity-based recommendation method.