Cost Estimates of Buildings' Floor Structural Frames with the Use of Support Vector Regression

Abstract

The paper presents some results of the research into a development of a cost estimating model that is capable of using information from building information model and implementing machine learning for cost prediction. Accurate estimates, provided throughout the whole construction project, allow for actual cost savings and assist in achieving sustainability goals. The model which is based on the support vector regression and radial basis kernel functions has been developed and proposed to support cost estimates of building's floor structural frames. The author's main assumption was to combine the benefits of building information modelling - namely the ability to extract certain information about the building and structural members of the floor frames from the models and the capabilities of machine learning. The research, presented in this paper, came down to solving a regression problem with the use of the support vectors approach. The training data for machine learning included inputs that represented features of the building and structural members' belonging to the floor structural frames and outputs that represented corresponding real life cost estimates of the floor structural frames. The obtained results show that the proposed model allows predicting costs with satisfactory accuracy.