I created a health risk ranking system that is 50% more accurate than commercial models. Technically, commercial models rank cost, not risk. My model ranks health levels.
I wanted to make the white papers available to anyone who is interested. I need to rewrite these, but I will err on the side of presenting OK, vs great.
I started by building a traditional health ranking model.
See the white paper here.
And the Technical Guide here.
Risk Model V25.9 Technical Guide
In 2017, I was able to get a 50% improvement in risk ranking accuracy through the use of a rank based inverse normal transformation.
See the explanation here.
Recent Comments