Research On Construction Cost Prediction Of High-Rise Residential Buildings Based On Stacking Algorithm

As a large-scale investment project,high-rise residential project has the characteristics of high risk,large scale and long construction period.How to find the basis for bidding quotation according to the characteristics of high-rise residential project itself has been the focus of construction enterprises.Because the traditional construction cost prediction method requires a lot of data and time support,it cannot provide a more accurate basis for the bidding price of construction companies in a relatively short period of time.Therefore,this paper proposes a prediction method based on the Stacking algorithm to solve the problem of construction companies.The problem of high-rise residential construction cost prediction in the case of small sample data provides a reference for actual projects.First,the cost structure of the high-rise residential project is combined with the actual situation to construct a complete model evaluation index system.The data is processed and analyzed for correlation and outliers first,and then PCA is used to reduce the dimensionality of the data.This method can not only retain the original data features,but also extract the features of the data to ensure the accuracy and accuracy of subsequent modeling and timeliness.Secondly,combined with the application situation and classification of the first-level model in machine learning,first use the default parameters to compare the prediction effects of the first-level model,select the three first-level models with accurate prediction results,and then combine the learning curve and grid search methods to compare one.The new training set and test set build the Stacking secondary model and get its prediction results.Finally,this article takes a real estate built high-rise residential project in Lanzhou as an example,and compares and analyzes the actual value of its final accounts,the predicted value of the first-level model.Finally,it is concluded that the Stacking algorithm has a better prediction effect for small sample data,which is better than that of other first-level models.This paper establishes a high-rise residential construction cost prediction model based on the Stacking algorithm.This model provides a fast and accurate construction cost prediction method for construction companies to bid for quotations.The prediction results show that the method provided in this article has shorter calculation time and higher prediction accuracy,which can meet the needs of actual projects.