Search

This article explores the linguistic features of different registers in Chinese through text clustering driven by the Menzerath–Altmann (MA) law. We propose to calculate the average word length distribution according to clause length. The MA law predicts that texts from different registers will show differences in terms of average word length distribution in texts. As predicted by the MA law, analysis result demonstrates that average word length decreases with the increase of clause length in each register and that their relationship can be fitted by the formula y = axb e - cx ⁠. We hypothesize that it is the situation type, i.e. whether the text is dialectic or monologue, that is the linguistic characteristic behind the dichotomy of word length distribution. To confirm these register-distinguishing linguistic features, texts were represented by the average word length distribution and the fitted parameters using the vector space model and clustered according to their register categories. Good clustering results show that average word length distribution in certain length clauses and their fitted parameters can be used as the distinctive characteristics of these three registers. [ABSTRACT FROM AUTHOR]

In this article, we propose an innovative and robust approach to stylometric analysis without annotation and leveraging lexical and sub-lexical information. In particular, we propose to leverage the phonological information of tones and rimes in Mandarin Chinese automatically extracted from unannotated texts. The texts from different authors were represented by tones, tone motifs, and word length motifs as well as rimes and rime motifs. Support vector machines and random forests were used to establish the text classification model for authorship attribution. From the results of the experiments, we conclude that the combination of bigrams of rimes, word-final rimes, and segment -final rimes can discriminate the texts from different authors effectively when using random forests to establish the classification model. This robust approach can in principle be applied to other languages with established phonological inventory of onset and rimes. [ABSTRACT FROM AUTHOR]