Can you make better estimations of regression parameters if you have censored data? Short answer, you do with Expectation Maximization.
When and how to use weigthed regression.
An example on over optimistic inference.
A few comments to have in mind about R2
Some details about OLS as smoothing of the training data. Weighted average of Y by distance of X to the center and variance of X.
Some details about Bias Variance Tradeoff
Basic usage of Pyspark and link to useful documentation.