12. Choosing the right estimator

Often the hardest part of solving a machine learning problem can be finding the right estimator for the job. Different estimators are better suited for different types of data and different problems.

The flowchart below is designed to give users a bit of a rough guide on how to approach problems with regard to which estimators to try on your data. Click on any estimator in the chart below to see its documentation. The Try next orange arrows are to be read as “if this estimator does not achieve the desired outcome, then follow the arrow and try the next one”. Use scroll wheel to zoom in and out, and click and drag to pan around. You can also download the chart: ml_map.svg.