Stylometric investigation of 19th-century English through text-mining

本研究尝试运用最先进的数据挖掘工具对19世纪英国散文进行文体学研究。在这项研究中分析的语言特征范围从非常常见的功能词到中低频项目。虽然19世纪的英语在原则上与18世纪的英语以及20世纪初的英语有着相似的语言特征，但事实证明，使用机器学习技术可以清楚地区分19世纪的文本与18世纪以及20世纪初的文本。通过机器学习挖掘算法提取的主要区别变量包括身体部位语言，其用作情感描述符或表征设备。更有趣的是，查尔斯狄更斯在世纪引发了这种文体手段的重大发展。

This grant-in-aid study has tried applying the state-of-art data-minig tools in stylistic investigation of nineteenth-century English prose. The linguistic features analysed in this research range from very common function words through mid- to lower-frequency items. While nineteenth-centurry English shares, in principle, similar linguistic features with eighteenth-century English as well as English in the early twentieth-century, it has proved to be possible to clearly differentiate nineteenth-century texts from texts written in the eighteenth as well as in the early twentieth centuries using machine learning techniques. Major distinguishing variables extracted though machine-learning mining algorithms include body-part languages, which function as descriptors of emotion or as device for characterisation. Of further interest is that Charles Dickens triggered the significant development of such stylistic devices in the nineteenth century.