Markers are internal labels that you can add to every sample to help you organize your training data. For example, each time you upload a file with training samples, each sample will be marked with a Marker with the title of the file, as well as a Marker with the name of the person who uploaded it.

When you upload training data using a CSV or Excel file, you can also select a specific column to be uploaded as Markers, again to help you organize and find data within the model.

Last but not least, all the text sent to be classified by the model, get a special marker named "inbox", which allows you to later use this to find the texts sent to classify.

Working with markers

You can access the markers through the Data section in the Build tab:

For example, the following model has training samples with two markers: '1st_dataset' and '2nd_dataset':

Markers are uniquely useful when exploring your training data; you can filter your data to view a specific subset of data labeled with a particular marker.

The following are the training samples filtered by the '2nd_dataset' marker:

You can use markers to identify different batches of training data (e.g. survey responses from Q1 vs Q2). Or you can use markers to label training data you uploaded to the model and differentiate it from the data added by a teammate.

Take into account that markers are entirely optional and are only for human purposes; markers are not used by MonkeyLearn to create or train a machine learning model.

Did this answer your question?