信息化浪潮极大地改变了人们的生产、工作、生活方式，也改变了科研工作人员的科研资料获取方式和信息沟通方式。传统的期刊评价体系通过揭示期刊间的相互引证关系，帮助研究人员在期刊的水平上测量研究成果的影响力。由于网络用户行为的可追踪性和可再次分析利用性，构建基于用户的电子期刊评价体系成为可能。本文基于网络日志数据和期刊引证关系，从五个维度（统计方式、时间、作者、期刊、用户使用行为）出发，形成一套基于用户使用行为的电子期刊评价体系。论文包括以下几部分：第一章是对论文研究意义、内容、总体框架的阐述。第二章针对网络环境下信息获取方式的改变，探讨了电子期刊评价现有指标的现状和优缺点，并对用户指标进行定义。第三章给出了基于用户使用行为的电子期刊评价体系以及五个维度的组合方式第四章在前一章的基础上，对电子期刊评价体系的每一个指标进行详细阐述，并分别从这些指标所需的各个基础首先提出了一套基于用户行为的电子期刊评价体系，然后分别阐述了二级指标和一级指标中各个指标的意义、计算方式等。在一级指标中，不仅有IF、CJR、H指数、特征因子等测度不同主体的指标，还利用日志数据提出了用户迁移指数、用户成熟度、用户新颖度等新指标。论文最后针对现阶段的研究状况和问题，就电子期刊评价体系实现进行了探讨，并对日志所需记录的数据提出建议。

Our daily activities have been deep changed by the infomationization, which also include researchers’ works. The traditional journal evaluation system is based on the journal citing relation to help researcher analyze their power of influence in their research field. With the traceability and reusability of the behavior of network user, it is possible to build a user behavior-based e-journal evaluation system. Based on log data and citing relation, this thesis build a user behavior-based e-journal evaluation system by discuss five dimensions which include statistic, time window, journal, user and user-behavior. Followings are what this thesis discussed. The chapter 1 discusses the reasons of why does this thesis have been wrote. The chapter 2 discusses the present situation and features of E-journal evaluation indexes, and defines the user index. A set of user behavior-based e-journal evaluation system and a table of five dimensions is proposed at chapter 3. The chapter 4 discusses meaning and formula of the first class index and the second class index separately. The first class index not only includes IF, CJR, H-index, Eigenfactor, but also includes some new indexes, such as User Migration Factor, the rank of user, click rate of user, downloads of journal, journal concern, and so on.. Then, this chapter discusses statistic indexes and theirs meaning from five dimensions. The final chapter discusses the creations and drawbacks in this thesis, and gives suggestions of contents that log data should include.