摘要
以同一题材但创作于不同历史时期的《三国志》(文言文)和《三国演义》(半文半白)为语料,以汉字在句子中从前字指向后字的相邻关系作为连线方向,分别为两个文本构建汉字相邻共现关系网络。研究发现两个不同时期的汉字网络具有相似的统计特征:高连接性的汉字使用频率较高;在整体上都具有小世界性、无标度性和协调非相关性;"分层云集性模型"可以部分解释汉字网络的三方谱特征,其中16-300和15-210这两种三方组在网络中具有优势分布。最后还讨论了两个网络的统计指标的差异及其意义。
This study used two texts from different historical periods, both of which concerned on the topic of Three Kingdoms Period. The Three Kingdoms had a classical style of writing and the Romance of the Three Kingdoms was a mixture of both classical and vernacular Chinese. By defi- ning immediate adjacent relations directed from former characters to later ones inside sentences as directed lines, two character adjacent co-occurrence networks were established for the two texts respectively. It was revealed that there are similar statistical properties embedded in these two networks of different periods: the highly connected characters have high frequencies; both of the networks have global patterns of small world, scale-free and disassortative mixing structures. Triad censuses of these two networks had suggested that two types of triad (16-300 and 15- 210) have prominent distributions and the "hierarchical clusters model" can partly explain the properties of their triad censuses. This paper also reported the differences between the statistical indices of the two networks and discussed their significance.
出处
《复杂系统与复杂性科学》
EI
CSCD
北大核心
2012年第3期50-61,共12页
Complex Systems and Complexity Science
基金
国家自然科学基金(81171854)
关键词
汉字网络
小世界性
无标度性
协调非相关性
三方谱
Key words: Chinese character networks
small world
scale-free
disassortative mixing
triad census