Leveraging Machine Learning to Predict IFC Project Success and Failure

This research expounds on the efficacy and impact of a machine learning-based approach in managing the investment services projects of the International Finance Corporation (IFC). It is achieved by focusing on the impact of financial interventions and project completion. We used a detailed dataset to assess trends in project status over time. The details about the growth of completed and active projects and the fluctuations in pending stages, like signing, disbursement, and approval are available in the dataset. We experimented with the ML models, including Random Forest, Gradient Boosting, Logistic Regression, SVC, and Decision Tree, to predict project outcomes. These models are evaluated with the metrics such as, accuracy, precision, recall, F1-score, and ROC AUC. After rigorous experiments, we observed that Gradient Boosting achieved the highest accuracy (90.76<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>) and Random Forest closely followed (90.43%). We identified significant trends that include a high success rate in project completion, an incremental growth in active projects, and a low occurrence of hold and pending approval statuses. We also performed statistical analysis and observed that loans dominate IFC's financial support and risk management and guarantees see less use. Additionally, correlations between total investments and other financial products highlight a diversified investment strategy. The findings emphasize opportunities for improving project approval efficiency, managing bottlenecks, and implementing data-driven decision-making to optimize investment strategies and enhance sustainable development outcomes. This research offers actionable insights for policymakers, demonstrating how ML can improve project management processes and maximize the impact of IFC investments in developing economies.