论文部分内容阅读
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。
Select high-frequency words for common-word network analysis will lose the information, select all the words in turn will make the common-word network structure confusion, can not be analyzed. In order to solve this problem, based on the E index, an improved method of calculating the total word strength is proposed. According to the co-occurrence frequency of words, the E index with higher co-occurrence frequency is given higher weight, co-occurrence frequency Low word pairs have a lower E-value. This method does not need to select high-frequency words, all words can be directly for the word network analysis. In order to verify the effectiveness of this method, an empirical comparison is made with the patent literature of electric vehicle battery as the data source.