Research On Construction Of Forecast Model Of Highway Engineering Cost In China

Since the opening of Shanghai-Jiaxing Expressway since 1988,the expressway in China mainland has been developing continuously for 30 years.At the beginning of14th Five-Year,the total mileage of China expressway has exceeded 5.1 million kilometres,and the number of expressway will keep rapid growth in the next few years.With the continuous growth of highway mileage,highway engineering cost index also presents an upward trend.With the development of society and the continuous growth of various macroeconomic indicators,there are great differences in the growth of highway engineering cost among provinces and cities.In view of this trend,China actively promotes the optimization and application of project cost management system,and integrates big data and project management informatization and other related technologies to predict and control expressway project budget.At present,the preparation of highway engineering cost estimate is still based on the national issued quota and the preliminary scale cost index database established by the relevant departments of various provinces.The preparation of highway engineering cost estimate still can not guarantee the reasonable and standardized use of huge investment.Therefore,it is particularly important to use the reasonable data regression model to predict the highway engineering cost index.By sorting out the factors affecting the highway cost index and using the form of literature review,case analysis,questionnaire and so on,this paper determines the project characteristic index and macro-economic index which affect the project cost index,and studies the correlation between various indicators for the highway cost index and various indicators.Subsequently,various forms of forecasting methods were used to construct the prediction model of the highway cost index.Among them,the Lasso regression prediction model with strong multi-collinearity processing ability and the XGBoost algorithm model which has been widely applied in the fields of finance and real estate were selected to construct the prediction model simultaneously.Through statistical analysis of 322 sets of data of provincial expressway construction projects in the five years from 2016 to 2020,RMSE,MAE,R~2 and other indicators of prediction results were used to evaluate the prediction ability of the two models,and the model with better prediction ability was selected to determine the parameter combination.Finally,through the powerful learning ability and prediction ability of XGBoost algorithm model,the paper analyzes the characteristic importance of each characteristic index in the cost index system,and sorts it,and forecasts and verifies the optimal parameter combination,and outputs the complete prediction model and correlation characteristic index of highway engineering cost index in China.