Title: Authorship Attribution of Poetic Texts Author: Mgr. Petr Plecháč, Ph.D. Department: Institute of Czech National Corpus Supervisor: doc. Mgr. Václav Cvrček, Ph.D. ABSTRACT Contemporary stylometry offers a number of methods for authorship recognition of po- etic texts based on a variety of textual features (e.g. word frequencies, frequencies of character n-grams). However, it seems that one important aspect of these texts has been rather left aside - this aspect is versification. The thesis uses four corpora of poetic texts (Czech, German, Spanish, and English) in order to analyze to what extent versification features - such as frequencies of rhythmic patterns or frequencies of various types of rhymes - may be used as an indicator of authorship. We show that (1) versification-based models significantly outperform the random baseline, (2) in some cases versification- based models even outperform the traditionally used lexical models, (3) in most of the cases combination of both types of models outperforms the given models alone. Versifi- cation features are consequently employed for the purpose of attribution of two texts of doubted authorship: (1) the versified play The Famous History of the Life of King Henry the Eigth which was originally published under the name of William Shakespeare, but where...

The aim of this thesis is to provide an overview of some of the commonly used methods in the area of authorship attribution (stylometry). The text begins with a recap of history from the end of the 19th century to present time and the required terminology from the field of text mining is presented and explained. What follows is a list of selected methods from the field of multidimensional statistics (principal components analysis, cluster analysis) and machine learning (Support Vector Machines, Naive Bayes) and their application as pertains to stylometrical problems, including several methods created specifically for use in this field (bootstrap consensus tree, contrast analysis). Finally these same methods are applied to a practical problem of authorship verification based on a corpus bulit from the works of four internet writers.