Learning classifiers without negative examples: A reduction approach

摘要: The problem of PU Learning, i.e., learning classifiers with positive and unlabelled examples (but not negative examples), is very important in information retrieval data mining. We address this through a novel approach: reducing it to the for some meaningful multivariate performance measures. In particular, we show how powerful machine algorithm, support vector machine, can be adapted solve problem. effectiveness efficiency proposed approach have been confirmed by our experiments on three real-world datasets.