Бірлескен тарату

Most probabilistic classifiers used for word-sense disam- biguation have either been based on only one contextual feature or have used a model that is simply assumed to characterize the interdependencies …...