Sobre la importancia de la reducción del espacio de búsqueda en la detección automática de plagio

In automatic plagiarism detection with reference, the text fragments in a suspicious document are exhaustively searched in a set of original (reference) documents in order to determine whether they have been plagiarised or not. One of the most important factors for the success of this kind of applications is the size of the reference corpus that, at the same time, may represent a problem when we consider performance and precision. In this paper, we approach automatic plagiarism detection analysing the impact of a preliminary search space reduction (composed of the original documents in the reference corpus). Our experiments over the METER corpus show that the Precision and Recall of the obtained results are improved when a search space reduction is applied at the beginning of a plagiarism detection process.