Working with Datasets

This guide covers how to manage your MarkovML Datasets

What happens when a dataset is registered with MarkovML?

When a dataset is registered with Markov, we use advanced algorithms to identify its data type, whether it is text, numeric, or categorical. We then apply unsupervised techniques that are specific to the data type to extract key insights. If the dataset is labeled, we also automatically train an AutoML model and evaluate the labeling quality using multiple label error detection algorithms. You can find more details about our approach [here].

First step

If you have datasets stored with a cloud provider like AWS, check out the Configure Cloud Access guide next.

Otherwise, proceed to register a dataset or data family.