Research On The Security Early Warning Method Of Building Construction Projects Based On PSO-SVM Model

The construction industry is the key industry of our national economy development, but its security risk and safety accident caused by the rapid development of the economic and social development has caused a great impact, how to improve the safety management level of the construction industry is a long time of research problems and research hotspot. The construction industry needs to improve the safety management mode, improve the level of safety management, and reduce the possibility of accidents.Exploration of the construction safety management mode of construction, the main is to get rid of the traditional management mode, the safety after the accident emergency and the responsibility, to the security warning before the accident, namely pre management. Effective security early warning is very important to reduce the occurrence of safety accidents and improve the safety management of building. It is very important to improve the objectivity and scientific nature of safety warning by using quantitative method. The traditional security early warning method uses the quantitative model in the parameters optimization, and can not get good warning effect. The particle swarm optimization(PSO) is superior to the traditional method in parameter optimization, and uses the information sharing between the whole group and individual to search for better parameters combination.In order to get better effect of building safety accident warning, this paper introduces the particle swarm optimization(PSO) based on support vector machine(SVM) prediction model, the main research contents include:First, the construction project safety warning theory, including the safety evaluation theory and early warning theory, the mechanism and the development process of the two kinds of theories, to lay the theoretical foundation for the following chapters.Second, the analysis of the construction projects in the presence of dangerous sources, and the risk sources and safety precautions as the starting point, according to the different characteristics of building safety prediction, based on the original safety evaluation system, building a new building safety evaluation system, so as to improve the safety and early warning data in the construction of the integrity and reliability.Third, in the optimization of the parameters, the algorithm has a significant effect on the optimization of the parameters. In this paper, the particle swarm optimization(PSO) is introduced, and the penalty factor and kernel function parameters in SVM model are optimized by PSO and PSO, and the PSO-SVM model is used to predict the safety of building construction projects by using SMV model.Fourth, using the traditional SVM forecast model and the improved PSO-SVM forecast model to carry on the empirical research to the construction project, the superiority of the PSO-SVM forecast model in the building safety early warning system is verified by comparison.Fifth, in view of the construction safety problems found in the process of building safety problems and suggestions, in order to effectively promote the construction project safety management level improvement and upgrading.