Research On Prediction Of Safety Accidents In Building Construction Based On Grey Neural Network

Today,the construction industry is one of the important industries in China,with a huge GDP and rapid growth rate,and has an unshakable position in our economy.Its safe development can make the national economy grow steadily and bring more employment opportunities to society,making the construction industry one of the hot industries in recent years.As the construction operation itself is random,dynamic and uncertain,considering the complexity in the construction process,this paper mainly establishes a safety accident prediction model with high accuracy.Based on this prediction model,it can provide a scientific basis for safety measures and build construction laws and regulations,and it can also formulate reasonable safety management goals.It is great significance to promote the sustainable and healthy development for the construction industry and society economy,contributing to the harmonious development of society.Firstly,analyzing the characteristics,types and occurrence rules of building safety accidents,finding out the main influencing factors leading to building safety accidents,and conducting grey correlation analysis on the main influencing factors.Since the original data of building construction safety accidents and their influencing factors are randomly oscillated,the original data should be subjected to accelerated translation transformation and weighted mean generation transformation to weaken its randomness.Then use the transformed data to establish a grey prediction model for construction safety accidents.It is concluded that the transformed grey prediction model is superior to the traditional grey prediction model.Finally,on the basis of analyzing the modeling advantages of both the BP neural network and the transformed gray prediction model,a combination model of multi-factor construction safety accident prediction based on the gray BP neural network is proposed,which is combined with the gray Markov model and a single model.The comparison of the gray prediction models proved that the accuracy of the method can effectively predict building safety accidents and provide a scientific theoretical basis for future building safety management.