Data processing of whole process cost management in building engineering based on AI

Abstract

Current housing construction projects are facing challenges such as cost overruns. Traditional cost forecasting methods can no longer meet the needs of the construction industry under digital transformation. This paper aims to explore a new cost forecasting method to improve forecasting accuracy and efficiency. This paper introduces the PSO-BP algorithm, which optimizes the weights and thresholds of BPNN to enhance the forecasting performance. The study selected 15 groups of housing construction project data with similar basic parameters as samples. After data cleaning, the BP neural network (BPNN) was used for cost prediction. To further improve the prediction effect, the particle swarm optimization (PSO) algorithm is introduced to form the PSO-BP algorithm to optimize the weights and thresholds of BPNN. Research results show that the PSO-BP algorithm performs better than BPNN in cost prediction. Specifically, when predicting the cost of Project 1, the error of BPNN is 7.67 yuan/square meter, while the error of the PSO-BP algorithm is only 2.12 yuan/square meter. In addition, when predicting the risks of 8 projects, the maximum error of BPNN is 0.18 points, while the maximum error of the PSO-BP algorithm is only 0.04 points. The results show that the PSO-BP algorithm not only improves the accuracy of cost forecasting but also significantly improves the efficiency and accuracy of risk assessment, helping to formulate more scientific cost control plans.