Trinary tools for continuously valued binary classifiers

Classification methods for binary (yes/no) tasks often produce a continuously valued score. Machine learning practitioners must perform model selection, calibration, discretization, performance assessment, tuning, and fairness assessment. Such involve examining classifier results, typically using summary statistics manual examination of details. In this paper, we provide an interactive visualization approach to support such continuously-valued tasks. Our addresses the three phases these tasks: operating point examination. We enhance standard views introduce task-specific so that they can be integrated into multi-view coordination (MVC) system. build on existing comparison-based approach, extending it continuous classifiers by treating values as trinary (positive, unsure, negative) even if will not ultimately use 3-way classification. cases demonstrate how our enables machine accomplish key