Hello Hasan, extraction - do you mean outliers, is this body fat measurements with callipers. Accuracy is how valid the measurements are - ie is your measuring instrument correct, calibrated and taking true readings, what is your comparison testing procedure? Reliability, do you get the same result if you have different operators or under different conditions? I'm not sure whether this is of help, please clarify pinch analysis. Thanks Debbie

Hi Hasan, i find the question stated to general to permit a precise answer -- which, i suppose, is what you hope for.

From what i understood, Pinch analysis is similar to Support Vector Classification, which makes it important to look at the data individually -- only hot and only cold -- before looking for the pinch point. Outliers is an important issue, like D. J. Hilton wrote already, but it is important to look at the outliers from different perspective: distribution, data quality and consistency, etc.