Classification Of Dangerous Chemicals Accident Cases Based On Convolutional Neural Network

With the development of the country and society,the hazardous chemical industry is booming.In recent years,due to various reasons,the accidents of hazardous chemicals have occurred frequently,which caused huge economic loss and casualties.So it has caused the state to attach great importance.The state issued a series of laws,regulations and standards during the 12 th Five-Year Plan period.Safety management methods and technical measures have been significantly improved.And the safe production situation of hazardous chemicals have continued to improve.In order to reduce the occurrence of hazardous chemical accidents and reduce the damage caused by hazardous chemical accidents,the main solution in this field is to establish the case database of hazardous chemical accidents.It can realize the information sharing of chemical accidents,which could help decision-making personnel to understand the accident and assist decision-making personnel to deal with similar accidents later.At present,the construction of hazardous chemical accident case database in our country lags behind that of western countries and have some problems such as simple structure,lack of case management mechanism,single retrieval method,and different case classification standards.With the increase of hazardous chemical accident cases and related news,case management method of the traditional hazardous chemical accident case database can no longer meet the needs of storing a large number of hazardous chemical accident cases.In view of the above problems,this paper attempts to conduct automatic classification research on a large number of online hazardous chemical accident case data,so as to strengthen case database functions of retrieval and management.Combined with the big data text classification technology in the field of natural language processing,this study uses the convolutional neural network model of deep learning to study the classification of hazardous chemical accident cases.The main research work includes:First of all,the case text preprocessing and word segmentation research.According to the characteristics of hazardous chemical accident cases,we conduct special text processing and stop word operation.In order to improve the effect of word segmentation,we use the dictionary supplemented with chemical words to conduct word segmentation operation on the accident case text.After that,the word vector of the case text is constructed.Based on Word2 Vec model,we construct the word vector representation of hazardous chemical accident cases to avoid the disadvantages of traditional text feature extraction.Finally,the classification model of the convolutional neural network with multiple convolution dimensions was constructed.Aiming at the particularity of case text,we designed a multi-convolution kernel neural network to classify hazardous chemical accident cases.And we designed a controlled trial and verified the advantages of the model.This study eventually trained a classification model for hazardous chemical accident cases with high accuracy,which can be applied to deal with simple classification of a large number of hazardous chemical accident cases.