Research On Macro-forecast Of Project Cost Based On Improved Grey Model ——Taking Guangdong Province As An Example

As the project of the cross-area domain is increasing,the cost management is dependent on the traditional pricing management method,which is unable to adapt to the market rhythm.Construction enterprises and real estate enterprises are increasing the demand for macro-engineering cost management in the region.High-rise residential is the main building type in the real estate market.It is the representative building that reflects the cost management level of the project.In recent years,the impact of the price increase of the construction materials,the influence of the policy regulation and the frequency of the project,the construction cost estimates,the risk of the large business of construction enterprises and real estate enterprises.Therefore,how to help enterprises to reduce the risk of regional project cost fluctuations effectively is an urgent problem for the industry.The construction cost index is an important indicator of the level of project cost in the area,and it has a very important effect on the cost control of the regional project of the enterprise.By analyzing the existing research and advanced cases of the industry,many scholars study the construction cost index of the whole country from a macro perspective,and use a single prediction model.However,the use of composite prediction models is relatively small for regional construction cost indices.So,based on previous studies,this thesis analyzes the key variables of the fluctuation of the cost index of the work process,and constructs the composite model for the prediction analysis on the basis of the multivariate grey prediction model.Firstly,this thesis summarizes the literature,and the three types of macro factors of the cost index of the work process,the economic condition,the construction market condition and the energy cost condition;Second,data from the national bureau of statistics,the digital database and other sources for the search and extraction of target data have been collected with 13 potential factors;Again,the raw data is processed by data mining technology,and the key factors of the construction cost index are selected by the combination of characteristics engineering.Then,this thesis combines the characteristics of the characteristics and determines the optimal characteristics variable set through the multivariate grey prediction model.Then,this thesis makes a prediction analysis of the improved grey model by combining the combination model.Finally,this thesis analyzes the accuracy of the empirical results of Shenzhen and Guangzhou,and verifies the advanced nature of the prediction accuracy of this research method.The results show that the three key factors of the construction cost index are the total amount of GDP,oil prices and consumer goods.The most important factor in the impact of the cost forecast is the economic situation.After the cross-validation of the sliding window of the multi-variable grey model,it not only obtains the smaller prediction error,but also has good linear prediction ability.Moreover,combined with GBDT's combination model,it can effectively improve the prediction accuracy and meet the actual use requirements relative to a single multi-grey model.In general,macro factors are the key factors of the project cost index,and the combination model can effectively improve the performance of the prediction model.This study comprehensively analyzes the macro influence factors of the construction cost index,and defines the important degree and optimal time of the various factors of the cost index,and builds the prediction model of the construction cost index through the combination model.The results of this study can provide reference and reference for construction enterprises and real estate in regional construction cost management and risk.