Beyond the Selected Completely At Random Assumption for Learning from Positive and Unlabeled Data
This addresses the challenge of biased data selection in machine learning for practitioners, though it is incremental by building on existing positive-unlabeled learning frameworks.
The paper tackles the problem of learning from positive and unlabeled data under selection biases, proposing a method that improves classifier performance by incorporating the labeling mechanism, with empirical results showing enhanced accuracy.
Most positive and unlabeled data is subject to selection biases. The labeled examples can, for example, be selected from the positive set because they are easier to obtain or more obviously positive. This paper investigates how learning can be ena BHbled in this setting. We propose and theoretically analyze an empirical-risk-based method for incorporating the labeling mechanism. Additionally, we investigate under which assumptions learning is possible when the labeling mechanism is not fully understood and propose a practical method to enable this. Our empirical analysis supports the theoretical results and shows that taking into account the possibility of a selection bias, even when the labeling mechanism is unknown, improves the trained classifiers.