Sorted Score Vector
The vector obtained by arranging the nonconformity scores of all calibration samples in ascending order. It is used to perform density ratio estimation in a low-dimensional space in order to avoid the curse of dimensionality.
Timeline (last 90 days)
That study proposes a new method for density ratio estimation based on ranked score vectors.
The researchers prove that, under interpretable conditions, the method based on ranked score vectors asymptotically recovers the same miscoverage rate report as the input-space estimate.
All Facts (2)
The researchers prove that, under interpretable conditions, the method based on ranked score vectors asymptotically recovers the same miscoverage rate report as the input-space estimate.
50%UnverifiedThat study proposes a new method for density ratio estimation based on ranked score vectors.
50%