A new data mining-based framework to predict the success of private participation in infrastructure projects

Abstract

The study aims to propose a data mining-based framework to predict the success of private participation in infrastructure projects in developing countries. Data have been collected from the World Bank’s maintained PPI projects database. The proposed framework in this study consists of imputation of missing values, selection of significant features method, resampling imbalanced classes, and application of classification algorithms, including random forest, logistic regression, and support vector machines to predict the binary classes (project success). The results suggest multivariate imputation by chained equations(MICE) as the best method for the imputation, Boruta for the feature selection method, and logistic regression for the classification to predict binary classes in PPI project dataset. The major contribution of this study is that it builds a new data mining-based framework, which considers different feature selection methods and classification techniques. This study will help the practitioners to predict the success of projects carried out under different contractual arrangements and adopt different proactive project management approaches.

Keywords:

• Logistic regression
• random forest
• support vector machine
• feature selection methods
• oversampling

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.