Spatio-temporal Semantic Analysis of Safety Production Accidents in Grain Depot based on Natural Language Processing

Abstract

At present, the grain industry mainly relies on the text information of production accidents entered by field personnel and expert experience to classify the causes of grain depot accidents. On the one hand, in order to accurately identify and classify the causes of grain depot accidents, security inspectors are required to have rich field work experience; On the other hand, the complicated and rare accident causes are limited by the experience of security personnel, which can easily lead to misclassification. Natural language processing (NLP) is an important direction in the field of artificial intelligence. It mainly uses an intelligent and efficient way to systematically analyze, understand and extract text data. Therefore, this paper puts forward the spatio-temporal semantic analysis of grain depot safety production accidents based on natural language processing, aiming at solving the problem of automatic classification of grain depot safety production accident text data. The text classification method based on word2vec and CNN not only considers the correlation between words, but also considers the relative position of words in the text, which has greater advantages than traditional feature selection methods. According to Word2Vec combined with CNN model, the grain accident text data is vectorized and feature extracted, which effectively improves the accuracy of grain big data classification.