Improving Project Budgeting Systems by Developing Machine Learning Models

Abstract

The lack of an efficient budgeting system makes it more difficult for a business to satisfactorily execute projects or gain new business. To improve the accuracy of budgeting using the classical approach, a dynamic system is required. Building dynamic systems that apply machine learning techniques can support companies in improving their budgeting system. This quantitative study built five machine learning regression models: multiple linear regression, artificial neural network, support vector machine, k-nearest neighbors, and random forest. The five built models were used to predict the closing costs of 552 industrial automation projects that were carried out in Africa and the Middle East. Using root mean square error, the model forecast precision was compared to that of the classical system. The outcome shows that there is a significant difference between machine learning models and classical systems. Therefore, the use of machine learning techniques can improve the accuracy for businesses of their budgeting system.