Positive-Unlabeled Classification under Class-Prior Shift: A Prior-invariant Approach Based on Density Ratio Estimation
This addresses a practical limitation in PU classification for applications where class-priors shift between training and test data, offering an incremental improvement over existing methods.
The paper tackles the problem of positive-unlabeled classification under class-prior shift, where training and test data have different ratios of positive samples, by proposing a prior-invariant method based on density ratio estimation that does not require class-priors during training. The result is a theoretically justified approach that demonstrates effectiveness in experiments.
Learning from positive and unlabeled (PU) data is an important problem in various applications. Most of the recent approaches for PU classification assume that the class-prior (the ratio of positive samples) in the training unlabeled dataset is identical to that of the test data, which does not hold in many practical cases. In addition, we usually do not know the class-priors of the training and test data, thus we have no clue on how to train a classifier without them. To address these problems, we propose a novel PU classification method based on density ratio estimation. A notable advantage of our proposed method is that it does not require the class-priors in the training phase; class-prior shift is incorporated only in the test phase. We theoretically justify our proposed method and experimentally demonstrate its effectiveness.