Modifications to p-Values of Conformal Predictors

Lars Carlsson, Ernst Ahlberg, Henrik Boström, Ulf Johansson, Henrik Linusson

Abstract

The original definition of a p-value in a conformal predictor can sometimes lead to too conservative prediction regions when the number of training or calibration examples is small. The situation can be improved by using a modification to define an approximate p-value. Two modified p-values are presented that converges to the original p-value as the number of training or calibration examples goes to infinity. Numerical experiments empirically support the use of a p-value we call the interpolated p-value for conformal prediction. The interpolated p-value seems to be producing prediction sets that have an error rate which corresponds well to the prescribed significance level.
BibTeX
@inproceedings{carlsson2015modifications,
  author = {Lars Carlsson and Ernst Ahlberg and Henrik Boström and Ulf Johansson and Henrik Linusson},
  title = {{Modifications to p-Values of Conformal Predictors}},
  booktitle = {SLDS},
  year = {2015},
  doi = {10.1007/978-3-319-17091-6\_20},
}

Loading related content...

Figure