No abstract

Burrows's Delta is the most established measure for stylometric difference in literary authorship attribution. Several improvements on the original Delta have been proposed. However, a recent empirical study showed that none of the proposed variants constitute a major improvement in terms of authorship attribution performance. With this paper, we try to improve our understanding of how and why these text distance measures work for authorship attribution. We evaluate the effects of standardization and vector normalization on the statistical distributions of features and the resulting text clustering quality. Furthermore, we explore supervised selection of discriminant words as a procedure for further improving authorship attribution.

We describe design and implementation of the linguistic query language DDDquery. This language aims at querying a large linguistic database storing a corpus of richly annotated historic German texts. We use a graph-based data model that supports multiple independent annotation layers on a shared text layer as well as alignments of text layers representing the same text or related texts (e.g., translations). The corpus is stored in an object-relational database system with a text retrieval extension. DDDquery is based on XPath to leverage the familiarity of many users with this language. It is translated to SQL in a two phase compilation with first order logic as an intermediate language. This approach effectively decouples the query language from the schema of the underlying corpus. We provide an overview of DDDquery, the underlying ODAG data model, its implementation as relational schema, the predicates of the intermediate language, and describe both phases of the translation process.

No abstract

No abstract