Unlocking business value from model interpretability

When it comes to predictive modeling, there is a trade-off between wanting to know what is predicted or why a prediction was made, or you do not care why a decision was made. Each of those scenarios depends on the use case. For example, if you are building a model for research and development (R&D) purposes, then you would sacrifice model interpretation over accuracy, whereas if you are building a model for business use, it is important to understand how the model is making that decision. Either way, it is important to know the model’s behavior not only to understand why some decisions were made but also for debugging and model improvement.

Let’s now understand why model interpretability for DL is important and how it helps to unlock practical business benefits such as better business decisions, building trust, and increasing profitability in the following section.

Better business decisions

Before model...