I also got the ESTPAB problem. When I tried running makeosmex to recompile mex files I got the error below.
xcodebuild: error: SDK "macosx10.7" cannot be located.
xcrun: error: unable to find utility "clang++", not a developer tool or in PATH
mex: compile of ' "estcondentropy.cpp"' failed.
Unable to complete successfully.
Error in makeosmex (line 25)
None of the solutions on this board worked for me (e.g. changing log(2) to log(2.0) or to log(double(2)), etc.
The problem turned out to be that I have Xcode 5.0 and Mac OS10.8, which apparently does not work out-of-the-box. The solution suggested by the Mathworks support team worked for me:
After that, makeosmex compiled for me and I was able to run demo_mi.
Other than that this package looks good.
Is there any assumption made when calculating mutualinfo(x, y) if one of vector x,y is binary?
In my data, y is binary and I tried several different x. Surprisingly, several different x give the same mutualinfo to the last digit of double type.
I think some assumption must be made on estimating kernel density. Could author give some insight?