An effective feature-weighting model for question classification

Question classification is one of the most important sub-tasks in Question Answering systems. Now question taxonomy is getting larger and more fine-grained for better answer generation, Many approaches to question classification have been proposed and achieve reasonable results. However all previous approaches use certain learning algorithm to learn a classifier from binary feature vectors, extracted from small size of labeled examples. In this paper we propose a feature-weighting model which assigns different weights to features instead of simple binary values. The main characteristic of this model is assigning more reasonable weight to features: these weights can be used to differentiate features each other according to their contribution to question classification. Furthermore, features are weighted depending on not only small labeled question collection but also large unlabeled question collection. Experimental results show that with. this new feature-weighting model the SVM-based classifier outperforms the one without it to some extent.