Forecasting IT Project Completion Time: Artificial Neural Networks Approach

This paper presents a methodology for predicting the required time for software development in IT projects and the necessary labor costs, based on trained artificial neural networks. The study aims to develop forecasting models to predict the execution time of IT projects and to improve project planning, control, and pricing methodologies in IT companies. Four models are developed for predicting execution time for specific types of IT projects, including simple and complicated learning models based on artificial neural networks. To validate the accuracy of the developed methodology, the study utilizes expert assessment of the technical management of the project. The results of the study demonstrate the effectiveness of the developed forecasting models, which have been successfully applied to plan and control projects with subsequent correction of results by retraining the developed models and the specifics of the client, country, and type of project.