Lecture 9: PCA MODELs of the Position_QB
PCA MODELs of the Position_QB. This video takes you into the 2020 QB metrics and using the PCA statistical technique provides a new metric for viewing QB last year’s activity. This sets the stage for a 2021 baseline (Reference Class Forcast see lesson 7_1)
I used PCA to reduce the complexity of data. It detects linear combinations of the input fields that can best capture the entire set of fields’ variance. The components are orthogonal to and not correlated with each other. The goal is to find a small number of derived fields (principal components) that effectively summarize the original input fields’ information.
I am an applied guy, as most biologists are. We use tools to address questions. If you are curious, I use JASP ( https://jasp-stats.org/ ) as a PCA program and other things, so what you see is their output. JASP defines each input as a variable to find the PCAs. PCA is really prep for serious regression. Think of it as a triage on your data sets. Then based on the PCA, you apply other techniques (Beyond Me 🙂 )
KEY Takehome is the idea of PCA is simple — reduce the number of variables of a data set, while preserving as much information as possible.
We have arrived at the “end of my fingertip knowledge”. I use the PCA to cut down my efforts to describe the past player performances. Do fewer variables mean less time for analysis? I hope! Hehe.
I have not done multiple or logistic regressions using these PCA. I think there are more aspects to be judged before regression analysis. Good Luck
Winning Your Fantasy Football Drafts Using the Professor’s Process. Lessons to Strengthen Your 2021 Fantasy Football Season. NFL Fantasy Football. Twitter @Prof_Fantasy1
Past Lesson Links