PREDICTING CONSTRUCTION PROJECT DURATION WITH SUPPORT VECTOR MACHINE

Abstract

This paper presents a forecasting model for construction time, using support vector machine (SVM)-recently one of the most accurate predictive models. Every construction contract contains project deadline as an essential element of the contract. The practice shows that a considerably present problem is that of non-compliance of the contracted and real construction time. It's often the case that construction time is determined arbitrarily. The produced dynamic plans are only of formal character and not a reflection of the real possibilities of the contractor. First, a linear regression model has been applied to the data for 75 objects, using Bromilow's "time cost" model. After that a support vector machine model to the same data was applied and significant improvement of the accuracy of the prediction was obtained.