Algorithm Selection


What is Algorithm Selection?

Supervised learning algorithms are essentially algorithms that approximate a function (f) that best maps input variables (x) to an output variable (y) from a given dataset [1]. As the field of machine learning advances, new and different types of algorithms are being researched and made available for training. Yet, there is no single best algorithm for all different types of data. Hence, algorithm selection is a key part of machine learning workflow.

The optimal type of algorithm(s) depends on the size of your dataset, structure, and the type of problem you are trying to solve. The best performing algorithms for a given dataset can be found via trial and error. Some of the common algorithms include linear regression, random forests,  gradient boosting, and deep neural networks [2].

Driverless AI supports a variety of algorithms, see resources below for more information. Additionally, Driverless AI allows you to customize your experiment by uploading a custom recipe of a model/algorithm of your choice. 





Related Links

No Related Resource entered.