Abstract: Based on a large Chinese journals database query log, this paper analyzed the basic features of academic database using behavior, including user access time distribution, special retrieval using proportion, as well as key words length distribution. Furthermore, the search strategy of different universities users was also analyzed. It is found that, according to users from different types of universities (science and engineering, normal, agriculture and forestry, medicine), there is a corresponding difference in search strategy; but as to users from universities of different educational levels (985 project, 211 project, etc), the differences in search strategies do not exist. Finally, by judging the IP address, the paper extracted users’ gender, profession, education level and other demographic variables. The authors analyzed the influence of these variables on their search behaviors, and proposed a factor model of university users’ search strategies.

Key words: log mining, search behavior, academic journals database, university users, human information behavior(HIB)

摘要： 基于国内某大型期刊数据库的用户查询日志, 探讨了高校用户使用学术期刊数据库的基本行为特征, 包括用户访问时间分布、使用特殊检索的比例、检索词的长度分布等。分析了不同类型高校用户的检索策略, 结果显示: 不同类型(理工类、师范类、农林类和医学类)高校用户的检索策略存在相应差异; 但不同办学层次(“985”高校、“211”高校等)高校用户的检索策略相差不大。最后, 通过IP数据的地址判断, 提取了高校用户的性别、专业和学历等人口统计学变量, 分析了这些变量对用户检索行为的影响, 提出高校用户检索策略的影响因素模型。

关键词: 日志挖掘, 检索行为, 学术期刊数据库, 高校用户, 人类信息行为