Strategic Predictions and Explanations By Machine Learning

Many machine learning (ML) models can make predictions regarding credit default swaps (CDS) for the telecommunication (telco) service sector. However, some algorithms can only offer a black-box model. It is crucial to explain the prediction result for strategic decisions. We study the current state-of-the-art by comparing various ML models, including deep learning (transformers), gradient boost machine (GBM), and extreme GBM (XGBM), plus various explanations tools, namely Variable Importance (VI) Partial Dependent Plots (PDP), Local Individual Conditional Expectation (LIME), Interpretable Model-agnostic Explanations (ICE), and Shapley values (SHAP) for the prediction model. To search for an optimal solution, we implement a hyperparameter search by leveraging High-Performance Computing (HPC). We aim to draw an optimal model for strategic CDS investment decisions. Our experiment results show that the XGBM provides the best solution with fewer constraints