1303 Predicting Cost Overrun Projects with Missing Data

Abstract

In this paper, we propose a method of predicting cost overrun projects with missing data. We use evaluated values of risk factors to predict the cost overrun. However, the risk factors often include missing values for various reasons. The missing values affect the accuracy of the prediction when using linear discriminant analysis. In this paper, we propose a method of predicting cost overrun projects using collaborative filtering, which is robust against a missing value. As a result of our experiment, the prediction accuracy of the proposed method is higher than linear discriminant analysis.