Development of construction duration prediction model for project planning phase of mixed-use buildings

Abstract

The nature of the mixed-use building’s characteristics should be considered to calculate a proper construction duration for mixed-use buildings in the project planning phase. However, the existing construction duration prediction models are incapable to reflect nature as they were developed mainly for single-use buildings. Therefore, the purpose of this study was to suggest a multiple linear regression model for mixed-use buildings that can help the owner and contractor predict a proper construction duration in the project planning phase. To do so, 1,969 mixed-use building project data were collected. The final prediction model for mixed-use buildings with five independent variables was suggested in this study. Its R2 was 0.637 and Ra2 was 0.635. In addition, the results of the validation showed that the prediction model’s average prediction accuracy, MAPE, and RMSE were 88.51%, 75.94, and 11.49% respectively. Given that the prediction model is applied in the project planning phase when there is insufficient information on the project, the results are believed to be at a considerably reliable level. This prediction model will allow the end-user to make simple and quick predictions of a construction duration of a mixed-use building based only on the basic information in the project planning phase.