Trinary Tools for Continuously Valued Binary Classifiers
Trinary Tools for Continuously Valued Binary Classifiers
复制标题
DOI：
10.1016/j.visinf.2022.04.002
复制
发表时间：
2022-04
期刊：
影响因子：
--
通讯作者：
Michael Gleicher;Xinyi Yu;Yuheng Chen
中科院分区：
文献类型：
--
作者：
Michael Gleicher;Xinyi Yu;Yuheng Chen
二元（是/否）任务的分类方法通常会产生一个连续值的分数。机器学习从业者必须执行模型选择、校准、离散化、性能评估、调优和公平性评估。这些任务涉及检查分类器结果，通常使用汇总统计和手动检查细节。在本文中，我们提供了一个交互式的可视化方法来支持这样的连续值分类器检查任务。我们的方法解决了这些任务的三个阶段：校准，工作点选择和检查。我们增强了标准视图并引入了特定于任务的视图，以便它们可以集成到多视图协调（MVC）系统中。我们建立在现有的基于比较的方法，将其扩展到连续分类器，将连续值视为三元（积极，不确定，消极），即使分类器最终不会使用3路分类。我们提供了一些用例，展示了我们的方法如何使机器学习从业者完成关键任务。
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 tasks involve examining classifier results, typically using summary statistics and manual examination of details. In this paper, we provide an interactive visualization approach to support such continuously-valued classifier examination tasks. Our approach addresses the three phases of these tasks: calibration, operating point selection, and examination. We enhance standard views and introduce task-specific views so that they can be integrated into a multi-view coordination (MVC) system. We build on an existing comparison-based approach, extending it to continuous classifiers by treating the continuous values as trinary (positive, unsure, negative) even if the classifier will not ultimately use the 3-way classification. We provide use cases that demonstrate how our approach enables machine learning practitioners to accomplish key tasks.