An Optimization of Machine Learning Approaches in the Forecasting of Global Financial Stability

Abstract

In the current data-driven world, the significance of machine learning as a mechanism for making predictions is vital. This research dives into how supervised learning techniques can be used to predict whether a banking crisis will occur in areas of Africa, which can be generalized to determining the status of financial stability in all areas around the world. By applying different machine learning mechanisms, along with tuning the hyperparameters, the optimal machine learning technique was found to be a neural network with two hidden layers, both hidden layers having the ReLU activation function. These results demonstrate that through widespread implementation of this neural network, governmental and financial organizations can develop significant trends and predict when a state is in economic peril, allowing for sufficient financial, social, or other aid to be administered before situations deteriorate.