共词网络分析中E指数的改进研究  

Studies on the Improvement of E-index for Co-word Network Analysis

作　　者：胡阿沛[1] 张静[1] 张晓宇[1] 

机构地区：[1]中国科学技术信息研究所,北京100038

出　　处：《情报理论与实践》2014年第1期46-50,共5页Information Studies:Theory & Application

摘　　要：选择高频词进行共词网络分析会遗失信息,选择所有词又会使得共词网络结构混乱,无法分析。为解决该问题,在E指数的基础之上,提出一种改进的共词强度计算方法,根据词对共现频次,赋予共现频次高的词对的E指数较高权值,共现频次低的词对的E指数较低权值。采用该方法无需选择高频词,可直接对所有词进行共词网络分析。为验证该方法的有效性,以电动汽车动力电池专利文献为数据源进行实证对比研究。Useful information would be lost if only high-frequency words are selected for co-word network analysis, while choosing all words would result in the confusion of the co-word network structure and be unable to analyze. To solve this problem, an improved method of computing co-word strength is proposed based on E-index. According to the co-occurrence frequency of the word pair, the E-index of the word pair of the high co-occurrence frequency is assigned higher weight, while the E-index of the word pair of the low frequency is assigned lower weight. By using this method, co-word network analysis can be di- rectly conducted on all words without choosing high frequency words. In order to verify the validity of the method, the paper makes an empirical and comparative study by using the patent literatures of the power batteries of the electric vehicles as the data source.

关 键 词：共词分析 E指数 共词网络 

分 类 号：G350[文化科学—情报学]