Unfortunately this script appears to be largely redundant given the built-in functionality of corrcoef.m. For example, I believe the following are equivalent:
nancorrcoef(A,B)
corrcoef(A,B,'rows','complete')
Moreover, unlike corrcoef this script only works with 2 variables - a serious limitation. Thus, though nicely written, there appears little reason to use this script. That said, it does have one unique feature in that it provides built-in checks/info on the proportion of NaN data. This is a nice idea, and perhaps the script could be re-written as a wrapper to corrcoef whilst retaining this convenient functionality.
Pros:
- Clearly written and well-documented
- Provides convenient checks on the proportion of NaN data
Cons:
- Only works with 2 variables
- Is largely redundant (can be done directly with corrcoef.m)
