Character-word Double-dimensional Semantic Classification Model for Judging Illegal and Irregular Behaviors for Internet Food Safety

摘要

It is difficult for Internet food safety supervision departments to automatically identify and classify the illegal and irregular behaviors in sampling inspection, an automatic semantic classification model was established to assist relevant departments in building an intelligent management system, so as to realize scientific decision-making. In this paper, Chinese word vectors and character vectors are used as the model input, CNN (Convolutional Neural Network) model is used to train the character vectors, and Positional Attention (PA) mechanism is introduced to BLSTM (bidirectional long short-term memory) network model to train the word vectors, to construct a character-word double-dimensional semantic classification model, CNN-PA-BLSTM, for judging illegal and irregular behaviors for internet food safety. The experimental results on the sampling inspection dataset show that the accuracy of the CNN-PA-BLSTM model is significantly higher than that of several commonly used deep neural network models, which verifies its rationality and effectiveness.