Research On Construction Waste Production Prediction Based On BP Neural Networ

With the advancement of urbanization,the continuous expansion of urban scale,large-scale housing construction and demolition activities brought the large increase of construction waste,environmental deterioration and ecological destruction more and more acute problems,bringing great pressure to our environmental protection work.Compared with developed countries,construction waste treatment started late,and was faced with construction waste production,recycling and low levels of problems.The reduction and recycling of construction waste has become an effective way to protect the ecological environment and the sustainable development of social economy.Therefore,how to scientifically and reasonably identify the influencing factors of construction waste production,forecast construction waste production and analyze its development trend is of great significance to promote the high-quality development of the city and build a"waste-free city"in an all-round way.Based on the influencing factors of construction waste production,this thesis predicts the output of construction waste based on BP neural network.The main research contents and results are as follows:The index system of influencing factors of construction waste production was established.Through literature reading,case analysis,research and other methods,13 indicators of influencing factors of construction waste production were selected from the four aspects of economy,management,society and humanity.SPSS software was used to carry out correlation analysis of the 13 indicators,and construct the index system of influencing factors of construction waste production.The main influencing factors of construction waste production were identified based on analytic hierarchy process.The judgment matrix of influencing factors of construction waste production was constructed by expert scoring.After normalization and consistency test of the matrix,the combined weight of each second-level index was obtained.The top 5 influencing factors with significant influence were screened out from the 13 index factors,which were as follows:Number of employees in the construction industry,labor productivity in the construction industry,recycling rate of solid waste,mechanization level of enterprises,total output value of the construction industry.The main influencing factors of construction waste production were identified based on grey relational degree model.The existing data matrix was determined as the comparison series and reference series,and the absolute difference between the corresponding elements of each index series and reference series was calculated.By analyzing the correlation coefficient and correlation degree of the series,and sorting the correlation degree,the main influencing factors of construction waste production output of the grey correlation degree model were obtained as follows:Number of employees in the construction industry,labor productivity in the construction industry,mechanization level of enterprises,per capita income of residents,recycling rate of solid waste.The output of construction waste is predicted based on BP neural network analysis.Based on the analysis of the results derived from the comprehensive analytic hierarchy process(AHP)and the grey relational degree model,the main influencing factors of the top 5 in the comprehensive ranking were obtained.Based on this,the prediction model of construction waste production was established and evaluated.The results show that the mean square error of the model is 1.906×10~8,the root mean square error is 0.230×10~4,and the prediction accuracy is0.87%,and the relative error every year is within±3%,reaching the model accuracy set.Input the data matrix of influencing factors of construction waste production,through the BP neural network model output results,construction waste production in 2022 will reach 2.79 billion tons,2023 construction waste production will exceed 2.8 billion tons,and in 2024,construction waste production will reach 2.86 billion tons.In this thesis,the index system of influencing factors of construction waste production is constructed to provide reference for construction waste management.The thesis constructs BP neural network prediction model,which enriches the prediction model system of construction waste production.At the same time,the output of construction waste from 2022 to 2024 is predicted to provide quantitative support for the reduction of construction waste.In this thesis,the government,enterprises and the public aspects of construction waste reduction to put forward targeted suggestions for construction waste reduction and resource management to provide references.