About this book
Machine learning (ML) interpretation enables practitioners to understand their models and mitigate risks associated with poor predictions.
The first section is a beginner's guide to interpretability, and it starts by recognizing its relevance in business and exploring its key aspects and challenges. It will reveal the inner workings of white-box models and contrast them to black-box and glass-box models and examine their trade-off. The second section is about mastering a vast array of interpretation methods while applying them to different use-cases. Model-agnostic methods studied range from permutation importance to SHAP and counterfactuals to sensitivity analysis. In addition to the step-by-step code, there's a strong focus on interpreting model outcomes in the context of each chapter's example. The third section is about tuning models and training data for interpretability by reducing complexity, mitigating bias, placing guardrails, and enhancing reliability. The methods explored here range from state-of-the-art feature selection and dataset debiasing methods to monotonic constraints and adversarial retraining.
By the end of this book, you will understand ML models better and enhance them through interpretability tuning.
- Publication date:
- February 2021