摘要: … As described previously, when training for a new author, we use n-fold crosstraining to determine what we expect to be good parameters. For each of the runs, we measure how well …
F.J. Tweedie, A.H. Neijt, R.H. Baayen, H. van Halteren, An experiment in authorship attributionMorin, A.; Sébillot, P. (ed.), Sixièmes Journées Internationales d'Analyse statistique de Données Textuelles. pp. 69- 76 ,(2002)