Research On Investment Estimation Method Of Prefabricated Building Based On BP Neural Network

At present,the prefabricated building has been vigorously promoted in China,and the investment amount of prefabricated building has been gradually increased.However,the cost of prefabricated building is higher than that of traditional cast-in-place building,and the investment is easily out of control.In the whole life cycle of a construction project,the investment decision-making stage has the greatest impact on the investment control of the project,and the early investment estimation of the assembled construction can provide an important basis for the feasibility study of the project and the selection of design schemes.The realization of the investment estimation can bring great convenience to the control and management of the cost of the assembled construction project.Therefore,it is of great significance to estimate the investment of prefabricated buildings in the investment decision-making stage.Firstly,this study introduces the theories and methods selected in this paper,and analyzes the advantages of BP neural network over traditional fast estimation methods.The classification of prefabricated buildings,the content of investment estimation,the requirement of precision and the function of investment estimation,and the concept of BP neural network are expounded in detail.It further establishes that BP neural network has great advantages over traditional rapid estimation methods.The reason why BP neural network is selected as the estimation model in this paper.Then,on the basis of reading a lot of relevant literature and combining with many years of prefabricated building engineering cases,the characteristic parameters that affect the cost of prefabricated building are determined,and the connotation of each characteristic parameter is defined.The main influencing parameters are screened by the screening evaluation model combining principal component analysis and expert scoring method.Then,the basic principle,network parameters,operation steps and MATLAB programming language of BP neural network are systematically elaborated,and the investment budget model of prefabricated building is constructed.Finally,typical projects from 2015 to 2018 are selected as training samples and prefabricated buildings of a primary school in changsha as test samples.Model accuracy is verified through simulation analysis of case projects.The verification results show that the prefabricated building investment prediction model based on BP neural network conforms to the investment prediction accuracy.The BP neural network prefabricated building investment estimation model constructed in this paper is based on MATLAB toolbox,running quickly and operating simply.It is verified by an engineering example,that it meets the precision requirement of investment decision-making stage and procides a high precison prediction method for investment estimation of prefabricated building.