For thesis exclusively
Add a plot of depolarization when talking about how difficult it is to model the actual angular dependency of depolarization. The code for this plot in correcting_P22_by_power_law.py
To improve the model
- Do not divide by xm to calculate coefficients. Instead divide based on \(2X_0\left|m-1\right|<<1\ \) which is one of the requirements of RDGA. This will most likely solve most of the complexity.
- If ever calculate the coefficients, then do not trust stepwise regression. It is misleading, too simple usually. Rely on Main effect plots, and interaction plots. Only include the variables if their plot makes sense. For instance in the following model: