Prefabricated Building Based On Adam-BP Neural Network Research On Construction Cost Forecast

Traditional buildings in my country’s construction industry have many problems such as large amount of consumables,material pollution and low production efficiency,while prefabricated buildings have the advantages of low carbon and environmental protection,fast construction,accuracy,safety and durability.In the face of my country’s environmental pollution,resource depletion,ecological destruction and other resource and environmental crises,since 2016,the State Council and the Ministry of Housing and Urban-Rural Development have issued a series of related policies to vigorously promote green and low-carbon buildings,and actively promote steel structures and prefabricated buildings.At present,my country’s construction industry has a low degree of standardization and industrialized construction,and the early investment cost of prefabricated buildings is too high,causing many developers to reluctant to choose prefabricated buildings.Therefore,solving the cost problem of prefabricated buildings will vigorously promote the development of prefabricated buildings in our country.The construction cost of a construction project greatly affects the profitability of the project.The accuracy of the construction cost prediction of a construction project is of great significance to the feasibility study of the project.In the project feasibility study stage,without construction drawings,engineers use their own project management experience to estimate costs and then make investment decisions.This kind of manual estimation is subjective,and the accuracy of the prediction result depends on the engineer’s existing engineering experience,which has limitations.Based on this,this article predicts the construction cost of prefabricated buildings,analyzes the factors that affect the construction cost,finds out the factors that are highly related to the construction cost,and proposes corresponding cost optimization countermeasures to provide a reference for the cost control of prefabricated buildings.This article firstly reads through the literature,expounds the domestic and foreign research status of prefabricated construction and engineering cost prediction,and introduces the machine learning and BP neural network model.Secondly,collect data of actual prefabricated construction projects,and establish a BP neural network prediction model based on Adam optimization algorithm.Screening and sorting out the collected data,establishing a training sample set containing 24 sets of data and a prediction sample set containing 4 sets of data,through training prediction,the prediction cost of the 4 sets of prediction samples is finally obtained.Through comparative analysis of actual construction costs and forecast costs,the results show that the Adam-BP neural network prediction model established in this paper has a good generalization effect and can accurately and quickly realize the construction cost prediction of prefabricated buildings.At last,Carry out a typical correlation analysis of 21 influencing factors that affect construction costs,and analyze the relative influence degree of each factor and construction cost.Choose an engineering example to comprehensively consider the results of the correlation analysis,and propose corresponding optimization schemes for the factors with large correlation influence to reduce the cost.The research in this paper makes the construction cost prediction results of prefabricated buildings more stable and more accurate,provides economic analysis and evaluation for investment decisions before bidding for engineering projects,determines the feasibility of the investment plan,and improves the economic benefits of the enterprise.