Expanding your machine learning toolkit: Randomized search, computational budgets, and new algorithms by Anonymous

Previously, we wrote about some common trade-offs in machine learning and the importance of tuning models to your specific dataset. We demonstrated how to tune a random forest classifier using grid search, and how cross-validation can help avoid overfitting when tuning hyperparameters (HPs). You'll learn a different strategy for traversing hyperparameter space - randomized search - and how to use it to tune two other classification algorithms - a support vector machine and a regularized logistic regression classifier. We'll keep working with the wine dataset, which contains chemical characteristics of wines of varying quality. As before, our goal is to try to predict a wine's quality from these features.