Summary
In this chapter, we learned about optimization techniques, especially the ones used in machine learning that aim to find the most effective hyperparameter configuration for an ML model fitted to a dataset. An optimized ML model has minimum errors, thereby improving the accuracy of predictions. There would be no learning or development of models without optimization.
We touched upon optimization algorithms that are used in operations research, as well as evolutionary algorithms that find usage in the optimization of deep learning models and network modeling of more complex problems.
In the final chapter of the book, we will learn about how standard techniques are selected to optimize ML models. Multiple optimal solutions may exist for a given problem and there may be multiple optimization techniques to arrive at them. Hence, it is essential to choose the technique carefully while building the model addressing the pertinent business question.
