Since we are debugging #1 , a possibly related problem is that the look-up tables are entered by hand, and are "magic parameters" in testfriedman.m . This is of course prone to bugs.
There is a way to approximate these tables and get rid of magic numbers: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5267387/
Since we are debugging #1 , a possibly related problem is that the look-up tables are entered by hand, and are "magic parameters" in testfriedman.m . This is of course prone to bugs.
There is a way to approximate these tables and get rid of magic numbers: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5267387/