摘要： 目的 提出面向期刊计量评价的多源数据融合方法,探析多源数据融合方法在期刊评价领域的应用价值,以期减少期刊评价中的信息缺失现象。方法 选取谷歌学术、百度学术、Web of Science、中国知网和维普数据库中共同收录的40种国内理工科期刊为样本,分析期刊计量结果的差异性,在此基础上提出多源数据融合方法,利用篇均被引频次进行实证研究。结果 不同数据源中的期刊计量评价结果差异较大。结论 综合内容权重和逻辑权重的多源数据融合方法不仅可以融合每种数据源中期刊计量结果的优劣,还可以平衡各种数据源中期刊排名的差距。

关键词: 多源数据融合, 期刊计量评价, 期刊篇均被引频次

Abstract: [Purposes] This paper proposes a multi-source data fusion method for journal measurement evaluation, and analyzes its application value, so as to reduce the lack of information in the field of journal evaluation. [Methods] We selected 40 Chinese journals of science and engineering jointly included in Google Scholar, Baidu Scholar, Web of Science, China National Knowledge Infrastructure, and VIP Database, and then analyzed the differences of journals measurement results. Afterwards, we put forward the method of multi-source data fusion. Finally, we used journal citations per paper to verify whether the method was reasonable or not. [Findings] The results of journal measurement evaluation in different data sources are very different. [Conclusions] Multi-source data fusion method combining content weight and logical weight can not only combine the pros and cons of data in each data source, but also balance the gaps in journal ranking among various data sources.

Key words: Multi-source data fusion, Journal measurement evaluation, Journal citations per paper