It's not clear to me that Halving Grid Search achieves better parameters than Randomized Search.

Alexandra Amidon

·Jul 14, 2021

…e can conclude that the test AUC-ROC score for the entire dataset (250000 sample) is almost the same. But the execution time of Randomized Search CV is 5x times faster and Halving Grid Search is 3x times faster than the Grid Search CV technique. For large-size datasets or param grids with a large number of components, it is recommended to use …

Stop using Grid Search Cross-Validation for Hyperparameter Tuning

Satyam Kumar

It's not clear to me that Halving Grid Search achieves better parameters than Randomized Search. The Test AUC-ROC Scores are roughly the same - RS is better for smaller sample sizes and HGS is better for larger sample sizes in the benchmark.

More from Alexandra Amidon

More From Medium

What’s Cooking! — Categorizing Cuisines with NLP

NLP-SENTIMENT ANALYSIS ON IMDB MOVIE DATASET

Particle filter

Supervised vs Unsupervised Learning

Content Based Image Retrieval Using Deep Learning

14 Popular Evaluation Metrics in Machine Learning

4 TYPES OF MACHINE LEARNING ALGORITHMS

Image-based ML techniques to classify billions of e-commerce products into thousands of categories