Hybrid principal component analysis and support vector machine model for predicting the cost performance of commercial building projects using pre-project planning variables

An accurate prediction of project performance in the pre-project planning stage – especially prediction of cost performance – is paramount to project stakeholders. The aim of this study is to propose and validate a hybrid predictive model for cost performance of commercial building projects using 64 variables related to the levels of definition in the pre-project planning stage. The proposed model integrates a support vector regression (SVR) model with principal component analysis (PCA). The proposed method was analyzed and validated based on 84 sets of data from an equal number of commercial building projects. Additionally, the result obtained using the proposed PCA–SVR model was compared with four other data-mining techniques. Experimental results revealed that the proposed PCA–SVR model is able to predict with high accuracy the cost performance of commercial building projects in the pre-project planning stage and is more efficient than the other four models.

► This study proposed a hybrid PCA–SVR model for cost performance prediction.
► The model was analyzed and validated on 84 commercial building projects.
► The model is able to predict the cost performance with high accuracy.
► The cost performance is dependent on definition levels in the pre-project planning.
► The model can be beneficial in developing early strategies and controlling project.