Prediction of Political Instability by Using Pre-Trained Neural Networks

This research aims to enhance and optimize the decision-making process in the political science domain by exploring the potential of machine learning. The aim was to create a pre-trained neural network to predict the political instability in any country (a prediction that aids decision-makers in handling government affairs and crisis prevention). We constructed four pre-trained neural networks, each tailored to a specific indicator (Human Development Index, Currency Strength Index, Tax to GDP Ratio, and Fragile States Index). These indicators are selected based on their strong correlation and how their concurrent performance impacts the political landscape of any country. The neural networks exhibited exceptional performance, achieving accuracy rates above 85%. The model built on the FSI demonstrated an astonishing accuracy of 99.67%, underscoring its potential for comprehensive assessments. The prospect envisions amalgamating the outputs of these pre-trained neural networks into a unified, deep-learning network, poised to yield collective decisions and recommend policy initiatives.