Abstract: Into the era of large data, information overload is a serious problem facing Internet users, personalized recommendation is a very promising way to solve this problem. In the academic field, academic resources personalized recommendation is an effective way to solve the information overload, which is recommended for users to meet their interest in personalized academic information. This paper discusses the recommendation of the existing academic resources from the perspective of user modeling, recommendation object modeling and recommendation strategy of personalized recommendation process. This paper summarizes the key points and existing problems of the research on the personalized recommendation methods of academic resources, including content recommendation, collaborative filtering recommendation and recommendation based on network structure, and the research trend of academic individuality recommendation is forecasted.

Key words: large data, information overload, academic resources personalized recommendation

摘要： 进入大数据时代，信息超载是互联网用户面临的一个严重的问题，个性化推荐是解决此问题的一个非常有潜力的办法。在学术领域，学术资源个性化推荐是解决信息超载的有效途径，其为用户推荐符合其兴趣的个性化学术信息。从个性化推荐过程的用户建模、推荐对象建模和推荐策略等三个模块角度对现有学术资源个性化推荐研究进行了探讨。针对目前广泛应用的学术资源个性化推荐方法，包括基于内容的推荐、协同过滤推荐和基于网络结构的推荐等，总结其研究的关键点和存在问题，并对学术资源个性化推荐的研究趋势进行了预测。

关键词: 大数据, 信息超载, 学术资源个性化推荐