Research On Development Of Construction Project Investment Estimation System Based On PSO-SVR

As the first step in investment control,the estimation of construction project investment has gained increasing attention.However,traditional methods of construction project investment estimation primarily rely on empirical estimation and manual calculations,which are insufficient to match the development of information technology in the construction industry and the need for precise cost control.Therefore,it is necessary to conduct intelligent estimation of construction project investment based on machine learning and database construction,and the research on investment estimation system construction serves as the foundation.Based on the analysis of the current status of construction project investment estimation,this study identifies several issues in current estimation practices,including incomplete consideration of estimation basis factors,challenges in information accumulation for estimation basis,significant influence of subjective experience in the estimation process,and inadequate information management in estimation practices.In response to these challenges,this study proposes the construction of a construction project investment estimation system consisting of two modules: the engineering data management module and the investment estimation prediction module.Regarding the engineering data management module,his paper identifies investment estimation indicators based on actual construction investment estimation needs and relevant research,selects the indicator system using the screening rules proposed in this paper,and applies the expert consultation method in the indicator selection process.A comprehensive investment estimation indicator system for construction projects is established.Additionally,the study addresses the standardization of engineering data input by considering commonly used BIM models and text data input methods,and designs a template for standardized engineering data collection.In the design of the investment estimation prediction module,the focus is on the prediction of engineering costs and other related expenses.Inspired by machine learning concepts,the study analyzes the preprocessing of sample data and establishes a PSO-SVR(Particle Swarm Optimization-Support Vector Regression)intelligent prediction model for engineering costs.Furthermore,the study achieves automated cost calculation based on fee rates.Finally,an empirical analysis is conducted based on a scientific office building construction project.The empirical case demonstrates that the proposed construction project investment estimation system is reasonable and feasible,simplifies the workflow of investment estimation compilation,and enhances the efficiency and quality of construction project investment estimation compilation work.