汉语句法网络的复杂性研究

旨在研究汉语句法网络的复杂性和复杂网络指标作为一种语言学研究手段的可行性,给出了如何按照依存句法树库构建句法网络的方法,自建了两种不同语体的汉语句法网络,对平均路径长度、聚集系数和度分布等复杂网络指标进行了测度.结果表明,两个句法网络和随机网络的平均路径长度和直径大致相当,但句法网络的聚集系数要远远大于随机网络,句法网络的度分布符合幂律,因此两种句法网络均为小世界和无尺度网络.发现两种语体的句法网络直径相同,但在平均度、平均路径长度、聚集系数和幂律指数方面则有较明显的差别.

The Complexity of Chinese Syntactic Network

The paper aims at the complexity of syntactic network and the feasibility that the complex network work as a means of linguistic studies. The paper proposes the method how to build a syntactic based on dependency treebank and investigates the complexity of Chinese syntactic dependency network based on two Chinese treebanks with different genres. The results show that syntactic networks have similar average path length and diameter with the random networks, but cluster coefficients of syntactic networks are much greater than that of random networks, and degree distributions of syntactic networks also obey the power law. The paper reveals that two syntactic networks have the same diameter, but with different average degree, path length, cluster coefficients and power exponent.