[MATH-598] Improving RandomDataTest

• Type: Improvement
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 3.2
• Assignee: Unassigned
• Reporter: Mikkel Meyer Andersen
• Created: 20/Jun/11 21:14
• Updated: 07/Apr/13 09:14
• Resolved: 17/Mar/13 17:19

Description

The tests is non-consistent and not good enough.

To quote psteitz from MATH-597:

Using testUtils.getQuartiles and chi-square is probably "best of breed." We might think about something fun like adding other quantile splits or even randomly generating quantiles (subject to bin size constraints).