Research On Measurement And Prediction Method Of Ground Material Information Price Based On PSO-RF

With the strengthening of environmental protection policies for rivers and mines,the production of ground materials such as sand and gravel has gradually decreased.However,the construction market has always had a huge demand for ground materials.The unbalanced relationship between supply and demand makes the price of ground materials continue to fluctuate greatly.In this context,the original ground information price release mechanism of the construction industry project cost management department has a serious time lag.The guidance for market economic activities such as project cost is not strong.It is also difficult to meet the higher requirements of the Ministry of Housing and Urban-Rural Development for the reform of project cost.Therefore,it is urgent to establish a scientific method for calculating the information price of ground materials and to accurately predict the fluctuation trend of the information price of ground materials.This paper studied the two core tasks of calculation and prediction of ground material information price,in order to promote the improvement of the ground material information price release mechanism.At present,the information prices of ground materials are mostly comprehensive prices without detailed analysis.It is easy to cause the problem of unclear price meaning or price distortion.First of all,the paper determined the price composition of the ground material information price through field research,and analyzed the influencing factors of the information price of ground materials.Based on that,it constructed the pricing system of land material information price.Then,according to different data sources and information collection,it determined the measurement methods of various indicators in the pricing system and carried out scientific measurement of ground material information price.Then,through literature analysis and comparative research,the particle swarm optimization-random forest(PSO-RF)algorithm had significant advantages in processing small sample and high-dimensional data.It can calculate the importance of influencing factors.And it was stable.Therefore,the PSO-RF method was chosen to predict the future multi-period information price of ground materials.In order to further improve the applicability of the prediction model,the paper established three prediction models by applying the PSO-RF method.The first prediction model focused on the information price data of the previous period.The second prediction model focused on the factors that affect the fluctuation of the information price of ground materials.The third prediction model simultaneously introduced the information price data and influencing factors of the previous period for prediction.This paper proposed a method that uses the second and third models to screen influencing factors and eliminate redundant influencing factor.It can sort the importance of the reserved influencing factors,and identify the main influencing factors that affect the fluctuation of the information price of ground materials.The three models can be selected reasonably according to different application requirements,application conditions and actual working conditions.Finally,this paper took the natural sand in the ground material as an example to verify.The actual data was collected to measure the price of ground materials and verify the applicability of the three prediction models.The research results show that the third prediction model that simultaneously introduce the data of the information price of the past period and the influencing factors has better performance.It can accurately reflect the future trend of ground material price changes.Its root mean square error RMSE is 0.5564,mean absolute error MAE is 0.4854,and mean absolute percentage error MAPE is 0.20%.At the same time,in order to verify the advantages of the PSO-RF model method,the third prediction model selected in this paper shows the optimal accuracy compared with other commonly used prediction models.The ground material information price measurement and prediction model determined in this paper can provide accurate reference for the calculation and forecasting work of the engineering cost department.It can effectively guide market economic activities and is conducive to optimizing the mechanism for releasing information on ground materials.