[Feature Request]: Rework Propensity Score Adjustments & add Calibration

[SvenKlaassen]

Describe the feature you want to propose or implement

I think it might be a good choice to rework

• ipw_normalization
• trimming_rule
• trimming_threshold

The goals would be to generalize reweighting schemes and other adjustments. If possible this should include the in_sample_normalization option of the doubleml.did module.

Further this should include an option to calibrate, see Algorithm 5 Calibration Strategies.

Propose a possible solution or implementation

• [ ] Include several functions in utils._propensity_score
• [ ] Rename trimming to clipping
• [ ] Include calbration of propensity scores