The authorship attribution process

Authorship attribution and stylometry, which refers to the statistical analysis of literary style, have always been closely related research fields. In fact, authorship analysis relies on the notion of style and on the process of drawing conclusions about authorship information of a document by analyzing and extracting its stylometric characteristics. This assumes that the author of a document has a specific style by which he/she can be completely or partly distinguished from another author. Following this idea, current authorship attribution methods have two key steps (see Figure 8.1):

• 1) an indexing step based on style markers is performed on the text using some natural language processing techniques, such as Part-Of-Speech tagging, parsing and morphological analysis;
• 2) an identification step is applied using the indexed markers to determine the most likely authorship.

An optional feature selection step can be employed between these two key steps to determine the most relevant markers. This selection step is done by performing some statistical measures such as mutual information or Chi- square testing [HOU 06].

The identification step involves using methods that fall mainly into two categories: the first category includes methods that are based on statistical analysis, such as principle component analysis [BUR 02] or linear discriminant analysis [STA 01]; the second category includes machine learning techniques, such as simple Markov chain [KHM 01], Bayesian networks, support vector machines (SVMs) [KOP 04] and neural networks [RAM 04]. SVMs, which have been used successfully in text categorization and in other classification tasks, have been shown to be the most effective attribution method [DIE 03]. This is due to the fact that SVMs are less sensitive to irrelevant features in terms of degradation in accuracy, and enable us to handle high-dimensional data instances more efficiently.

Figure 8.1. Prototype of the process of authorship attribution process