Choosing the Right Model for Your Use Case

So far, we have explored a set of white-box models and a couple of black-box machine learning models for the same classification use case. We also extended the same use case with a deep neural network in Keras and studied its performance. With the results from several models and various iterations, we need to decide which model would be the best for a classification use case. There isn't a simple and straightforward answer to this. In a more general sense, we can say that the best model would be a Random Forest or XGBoost for most use cases. However, this is not true for all types of data. There will be numerous scenarios where ensemble modeling may not be the right fit and a linear model would outperform it and vice versa. In most experiments conducted by data scientists for classification use cases, the approach would be an exploratory and iterative one. There is no one-size-fits-all model in machine learning. The process of designing and training...