I am currently studying on this site.
If I understand the flow of the story simply, I understood that this chapter is comparing models.
The content flow is
1-Use resampling to compare the four models.
2-However, the performance of the models will increase due to data dependency.
The rise and fall is proportional to any model, so it is possible that the performance is only increasing on average due to the way the data is divided.
3-To check this statistically, we will perform ANOVA.
4-To understand it more sensitively, we will use tidyposterior to obtain the posterior distribution of model performance.
This is how I understand it.
Are there any books or websites that can help me understand the mathematics of how the posterior distribution is obtained from the performance by resampling?
Is it inferior to tidyposterior to get multiple performance values by specifying a large value for repeat in vfold or using bootstrap to get a distribution of performance values?