An Artificial Neural Network Model to Estimate the Success Rate of a Project Based on key Success Factors

Abstract

Projects play a crucial role in identifying economic trends and guiding strategic decisions in project-based organizations. However, due to limited resources, uncertainty, and environmental complexity, investing in projects has become risky. Since project success is the ultimate goal of companies, identifying the essential factors that lead to success is of particular importance. The aim of this study is to present a model for early project evaluation and success prediction, as a risk analysis technique, based on the identified success factors. For this purpose, 120 projects from the Iranian Gas Engineering and Development Company, a subsidiary of the Ministry of Oil, were collected. Primary data were gathered through structured interviews with a group of experts (typically 8 to 12 individuals), and the success of each project was evaluated by this group using a seven-point Likert scale (ranging from "completely" to "fully"). A feedforward neural network approach was then employed to examine the relationship between critical success factors (CSFs) and project success. The MATLAB (R2016b, v 9.1.0) was used to develop the ANN model and create a graphical user interface (GUI). The results showed that the model's performance, with a test , was very good and demonstrated strong generalization capability. The model's accuracy was also considered acceptable from the experts' perspective. In fact, the model is highly effective in predicting project success (based on the experience of project managers) and can be used as a practical tool for risk analysis to assist managers in making timely and appropriate decisions