On Modeling and Querying of Text Corpora

This article proposes a novel data model for text corpora and discusses the issues on corpus query. First, a formalized definition of the corpus data is presented. Second, a data model is proposed in terms of the relational model, which is also proved to be complete. On this basis, we extend the query semantics of the traditional corpus query that generates KWIC (Keyword in Context) concordances and define the query problems. Finally, we investigate the data complexity of these querying problems and an experiment is also presented. These conclusions lay a theoretical foundation for the study of the modeling and querying of text corpora.