Multivatiate Approaches to Dickens' Style

方法问题的检查：从目前和过去的研究项目中收集和数字化的电子文本语料库中编制了各种类型的词频概况。数据类型包括高频、中频和低频词变量。一些词类也被用作变量，如as-ly副词和最高级。通过对应分析，本研究揭示了文本在作者之间的明显差异，以及文本在作者集合内如何根据时间顺序聚类。本研究中展示的多变量方法允许在多维空间中以图形方式检查文本之间的复杂相互关系，语言变量之间的相互关系以及文本与语言变量之间的关联。它以一种排列原始数据矩阵的方式计算行坐标（文本分数）和列坐标（单词分数），以便最大化单词变量和文本配置文件之间的相关性。在排列数据矩阵中，具有相似分布模式的语言学变量是最近的邻居，具有相似概况的文本也是如此。当文本/变量分数在多维图中投影时，变量条目之间的相对距离表示它们之间的亲和力、相似性、关联和其他类似关系。研究狄更斯风格的多元方法：从这种方法中获得的最有力的解决方案清楚地区分了狄更斯和控制集。特别有趣的是，作者的差异几乎出现在任何实验中，而不管使用的变量是什么。同样引人注目的是，从狄更斯的早期作品到后期作品，在大多数多变量分析中都发现了差异。更令人感兴趣的是，在狄更斯的作品中，系列小说的文本与小品和历史等其他类型的文本形成鲜明对比。这一系列结果说明了作者差异、文本体裁和年代是如何反映在狄更斯和对照作者所写文本的语言变量的频率模式上的。研究结果还表明，多变量方法在文本文体研究中的有效性。少

Examination of methodological issues : Various types of word frequency profiles have been compiled from the electronic corpus of texts which were collected and digitized in the present and the past research projects. The types of data include high-, mid-, and low-frequency word variables. Several word-classes are also used as variables, such as-ly adverbs and superlatives. By applying correspondence analysis, the present research has revealed how sharply texts are differentiated between authors as well as how texts are clustered according to chronology within authorial sets. The multivariate approach demonstrated in this research allows examination of the complex interrelationships among texts, interrelationships among linguistic variables, and association between the texts and linguistic variables graphically in a multidimensional space. It computes the row coordinates (text scores) and column coordinates (word scores) in a way that permutes the original data matrix so that the correl … More ation between the word variables and text profiles are maximized. In a permuted data matrix, linguistics variables with a similar pattern of distribution make the closest neighbours, and so do texts of similar profile. When the text/variable scores are projected in multi-dimensional charts, relative distance between variable entries indicates affinity, similarity, association, and other similar relationships between them.Multivariate approaches to the study of Dickens' style : The most powerful solution obtained from this methodology clearly differentiates between the Dickens and control sets. Of special interest is that the authorial differentiations have emerged in almost any experiments conducted, regardless of the variables employed. What is also striking is that a variation from Dickens' early works to later works was identified in most of multivariate analyses. Of further interest is that, in the Dickensian set, serial fiction texts stand in marked contrast to texts in other genres, such as sketches and history. The series of results illustrate how the authorial difference, text genre, and chronology are reflected in the frequency pattern of linguistic variables in the texts written by Dickens and the control authors. The results also suggest the effectiveness of the multivariate approach to stylistic studies of texts. Less