A language independent method for question classification

Abstract

Abstractsmall Previous works on question classification are based on complex natural language processing techniques: named entity extractors, parsers, chunkers, etc. While these approaches have proven to be effective they have the disadvantage of being targeted to a particular language. We present here a simple approach that exploits lexical features and Internet to train a classifier, in particular a Support Vector Machine. The main feature of this method is that it can be applied to different languages without requiring major adaptation changes. Experimental results of this method on English, Italian and Spanish show that this approach can be a practical tool for question answering systems reaching classification accuracy as high as 88.92%.

Question Classification in Spanish and Portuguese

“…In previous work we presented a language independent method for question classification were evaluation was performed on three languages: English, Spanish and Italian [2]. Although we achieved high accuracies we believe that considerable improvements can be attained by modifying some of the weakest features of this method, namely the set of heuristics chosen in order to construct the Internet queries.…”

“…These were the heuristics used in previous experiments [2], in addition to these, in this work we run queries eliminating words in the reverse direction. That is, if no hits are returned after eliminating the stop words, we eliminate the first word to the left and continue repeating this process until we have results.…”

A semantic approach for question classification using WordNet and Wikipedia

“…Solorio et al (2004) proposed an approach to question classification with only surface text and simple retrieval results from Google search engine. David et al (2006) proposed an automatic feature extraction approach to question classification, which uses only statistical information from unlabeled corpus to extract features without the help of natural language processing techniques.…”

Community Engagement and Quality Knowledge with Stackoverflow’s Reputation System: Applications to Bullipedia

“…SVM has shown good performance for many natural language related applications, such as text classification (Joachims 2002), and has been used in multiple studies relating to question classification (Blooma et al 2008;Tamura, Takamura and Okumura 2005;Solorio et al 2004;Zhang and Lee 2003). Table 4 summarizes the performance of all the classifiers evaluated in our experiment.…”