基于卷积神经网络的中文自动问答研究

iming at solving the problems of poor performance in chinese questioning answering system when using traditional rule-based or machine learning algorithm, a new deepging learning model was proposed to deal with chinese question answering system in this paper. Using deep structure of convolutional neural network, we measure the similarity of entity mentions with entities in the knowledge base and the similarity of relation patterns and realtions in the knowledge base. We score relational triples in the knowledge base using these measures and select the top scoring relational triple to answering the question. In this paper, we focus on processing fact-based questions and our method achieves higher precision.