Query Classification Using Convolutional Neural Networks

Nowadays, search engines become more and more important for people, it can help us to obtain the useful information. Capturing the query intent from use's query can improve the search efficiency and provide them a better user experiment. Query classification task can help us identify query intent. In this paper, we utilize convolutional neural networks for query classification, it is a supervise learning method, in the training process, we input the query and corresponding category to train this model, and in test process this model can predict query's category. In our experiments, we use two different convolutional neural networks to learn the query representation automatically based on the semantically similarity. Experimental results show that our model is competitive and can better predict query's category, the precision of one model improve 3% compared to logistic regression.