Fair Classification with Partial Feedback: An Exploration-Based Data-Collection Approach

arXiv:2402.11338v1 Announce Type: new
Abstract: In many predictive contexts (e.g., credit lending), true outcomes are only observed for samples that were positively classified in the past. These past observations, in turn, form training datasets for classifiers that make future predictions. However, such training datasets lack information about the outcomes of samples that were (incorrectly) negatively classified in the past and can lead to erroneous classifiers. We present an approach that trains a classifier using available data and comes with a …

abstract arxiv classification classifiers collection credit cs.ai cs.cy cs.lg data datasets exploration fair feedback form future information lending predictions predictive samples stat.ml training true type