Question classification using support vector machines

By Dell Zhang, Wee Sun Lee. Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, 2003

695 citations recorded by the source corpus.

Question classification is very important for question answering. This paper presents our research work on automatic question classification through machine learning approaches. We have experimented with five machine learning algorithms: Nearest Neighbors (NN), Naive Bayes (NB), Decision Tree (DT), Sparse Network of Winnows (SNoW), and Support Vector Machines (SVM) using two kinds of features: bag-of-words and bag-of-ngrams. The experiment results show that with only surface text features the SVM outperforms the other four methods for this task. Further, we propose to use a special kernel function called the tree kernel to enable the SVM to take advantage of the syntactic structures of questions. We describe how the tree kernel can be computed efficiently by dynamic programming. The performance of our approach is promising, when tested on the questions from the TREC QA track.

Subfields: Artificial Intelligence

View Paper: https://www.semanticscholar.org/paper/08745f22d0abbe66e486f0985c985ecf1eab4e9e

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Making Tree Kernels Practical for Natural Language Learning (2006, background): “For this purpose, it is enough to add the condition: 0. if n1 and n2 are leaves and their associated symbols are equal then ∆(n1, n2) = 1, to the recursive rule set for the ∆ evaluation (Zhang and Lee, 2003).”
Latent Support Measure Machines for Bag-of-Words Data Classification (2014, background): “, natural language processing [2], information retrieval [3, 4] and data mining [5].”
Efficient Convolution Kernels for Dependency and Constituent Syntactic Trees (2006, methodology): “Nevertheless, when the SST kernel applied to the dependency structures is combined with BOW, the SVM accuracy reaches 90% as well [16].”
Structured Lexical Similarity via Convolution Kernels on Dependency Trees (2011, methodology): “…that when no similarity is applied: (i) BOW produces high accuracy, i.e. 88.8% but it is improved by STK (the current state-of-the-art 5 in QC (Zhang and Lee, 2003; Moschitti et al., 2007)); (ii) PTK applied to the same tree of STK produces a slightly lower value (non-statistically signiﬁcant…”
Syntactic and Semantic Kernels for Short Text Pair Categorization (2009, methodology): “(Zhang and Lee, 2003), have shown accuracy comparable to the best models, e.”
Question Classification using Head Words and their Hypernyms (2008, background): “Zhang and Lee (2003) used linear SVMs with all possible question word grams, and obtained accuracy of 79 .”
From symbolic to sub-symbolic information in question classification (2011, methodology): “Further improvements of the tree kernel proposed by (Zhang and Lee 2003) are described in ( Pan et al. 2008 ), where semantic features (resulting from a set of 20 semantic classes such as PERSON, LOCATION and CREATURE) are incorporated into the kernel.”
Retrieving and Reading: A Comprehensive Survey on Open-domain Question Answering (2021, background): “For instance, [58] proposed to extract relevant words from a given question and then classify the question based on rules associating these words to concepts; [56] trained a list of question classifiers using various machine learning techniques such as Support Vector Machines (SVM), Nearest Neighbor”
Learning question classifiers: the role of semantic information (2005, methodology): “More recent works, including Radev, Fan, Qi, Wu and Grewal (2002), Harcioglu and Ward (2003) and Zhang and Lee (2003), exploit more involved machine learning techniques.”
Twitter Sentiment Analysis (2015, background): “It has been shown that the use of n-grams can improve the quality of text classiﬁcation (Raskutti et al., 2001; Zhang, 2003; Diederich et al., 2003), however there is no unique solution for the the size of n-gram.”
A survey on question answering technology from an information retrieval perspective (2011, methodology): “The success factor of SVM lies in introducing a syntactic or semantic kernel function [152,91].”
Question Classification by Weighted Combination of Lexical, Syntactic and Semantic Features (2011, methodology): “shown to have good performance on high dimensional data and generally outperform other classifiers, such as Nearest Neighbor, Naive Bayes, Decision Trees, SNoW and Maximum Entropy on question classification [16,6,5].”
Learning Answer-Entailing Structures for Machine Comprehension (2015, background): “Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing, pages 239–249, Beijing, China, July 26-31, 2015. c©2015 Association for Computational Linguistics”
Semantic Kernels for Text Classification Based on Topological Measures of Feature Similarity (2006, methodology): “Experiments on the TREC Question Classification Dataset The long tradition of Question Answering (QA) in TREC has produced a large question set which has recently been used for Question Classification (QC) tasks [9].”