Randomly remove data

[TTitcombe]

Feature Description

Update the data partitioning function to randomly remove datapoints from one (or both) of the datasets.

The chance of this happening should be small so that there is enough data left over on which to train the model

Is your feature request related to a problem?

In a real-world setting, we cannot guarantee that all datasets hold data relating to the same subjects. We wish to simulate that

Additional Context

Part of Epic #2
Dependent on #4