Open AccessPredicting in BIM Labour Cost with a hybrid approach Simple Linear Regression and Random Forest
Author(s) -
Jun Ze Wang,
Jie Zhang
Publication year - 2020
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/565/1/012108
Subject(s) - simple (philosophy) , simple linear regression , random forest , linear regression , simple random sample , computer science , proper linear model , regression , cost estimate , regression analysis , econometrics , engineering , statistics , mathematics , bayesian multivariate linear regression , machine learning , systems engineering , population , philosophy , demography , epistemology , sociology
In most cases, BIM labour cost can only be calculated through the proportion of the gross floor area according to the practical projects, which always brings unpredictable risks to project managers since this simple method of linear regression contains a high risk of estimating error in construction phases. Therefore, this study will develop a new hybrid methodology which combines the Random Forest and Simple Linear Regression for eliminating the error of prediction on BIM labour cost in construction phases. A case study will be conducted to illustrate the prediction results through four completed projects, including two steel structure and two reinforced concrete projects.