Big Data testing

[zachmayer]

This is a placeholder for Someone (myself or a volunteer) to do small project to test and improve caretEnsemble with large datasets.

Functions to test:

• caretList
• caretStack
• caretEnsembles

At least 3 test cases:

• Tall data: 1,000,000+ rows
• Wide data: 10,000+ columns
• Many models:caretList of 1,000+ models
• Others optional

These tests should be run via a script stored somewhere in this repo, and the data should be added via github lfs. The test results should be analyzed to identify bottlenecks in:

• RAM
• run time

Based on those results, we may do things like e.g. replace do.call, or use data.table in more places, or trim more data out of the model object, but it is premature to decide what to do until we've done some analysis

See also:
#155
#81
#70