Classification of Natural Language Semantic Relations under Deep Learning

Abstract

In order to accurately classify the semantic relations of natural language, the classification method of natural language semantic relations based on gated convolutional neural network is stated. The classification model of gated convolutional neural network is designed. The model is improved by combining dependent syntactic analysis and other external features. The analysis is carried out by calculating the macro average F1 value of the benchmark data set. The results show that the gated convolutional neural network model proposed in this study has the best classification results for evaluation using SemEval data set under different dimensional word vectors. In conclusion, the model proposed in this study can make full use of the depth of network model to improve the classification effect of semantic relations on small sample data sets. It is better to use models to classify semantic relations when incorporating multiple external features. This study provides an important reference for the application of deep learning in the classification of semantic relations.