3.1 What is BADA?

Purpose:

BADA - essentially perform PCA on Group Means (Barycenters) of a qualitative dataset. It is working like a unsupervised ML algo in which it classifies observations. BADA first creates a new space using Group Means, then projects the observations into that same space to see which classification would fit best.

Notes on Some Important Aids:

1. Heat Map: This is essentially a correlation plot of the averages of data with rows corresponding to the group and columns corresponding to the variables.
   
2. Fixed / Random - effect confusion matrices: This plot helps us measure accuracy of BADA by comparing Actual VS Predicted results from the BADA.
   

Author’s Notes:

BADA is the “Supervised” Version of PCA and thus can benefits from avoiding overfitting or bias during training and improves in classification performance. The process to do BADA is relatively easy given results from PCA. The Mathematical ideas behind BADA is very similar to PCA and only differs in that it uses Group Means for dataset. In short, BADA is like performing PCA on the group means.