This current has:
- a running (online) version for basic mean and variance.
- a running version with exponential weighting.
- a running correlation.
This is almost a straight rip from Statistics class oflthe python runstats package by Grant Jenks:
https://github.com/grantjenks/python-runstats
It also includes some other running statistics class from the GNU Statistics Library:
https://www.gnu.org/software/gsl/
I want to add the Regression class from runstats and some other algorithms from GSL that I need.
This untested code with no optimization thought. Hopefully that will improve, but my priorities shifted a little. Help is welcome.