| Título |
Complete syntactic N-grams as style markers for authorship attribution |
| Tipo |
Congreso |
| Sub-tipo |
SCOPUS |
| Descripción |
Human-Inspired Computing and its Applications 13th Mexican International Conference on Artificial Intelligence, MICAI 2014 |
| Resumen |
In this paper we present an authorship attribution method based on the use of complete (non-continuous, with bifurcations) syntactic n-grams as style markers. Syntactic n-grams are obtained by following paths in subtrees of a syntactic tree. We work with relatively short text fragments and build authors’ profiles of various sizes using tf-idf scheme. We train SVM classifier to perform the task. We compare the method with the application of character n-grams and show that the accuracy increases when using complete syntactic n-grams. |
| Observaciones |
Lecture Notes in Artificial Intelligence (including subseries of Lecture Notes in Computer Science) http://link.springer.com/chapter/10.1007/978-3-319-13647-9_2 ** Drive: Complete-syntactic_2014 |
| Lugar |
Tuxtla Gutierrez, Chiapas |
| País |
Mexico |
| No. de páginas |
9-17 |
| Vol. / Cap. |
8856 |
| Inicio |
2014-11-16 |
| Fin |
|
| ISBN/ISSN |
|
