From its excerpt:
Calibration can refer to predicting an explanatory variable from an observation of the dependent variable; to adjustment of measurements to agree with value of some standard; to transform classifier scores into class membership probabilities; etc.
These are quite unrelated uses sharing the same tag, only for language reasons. Especially the first mentioned use, inverse prediction, is quite unrelated to other uses. Should we split this and introduce a new tag 'inverse-prediction' (or in some other way)?