What to Expect of Classifiers? Reasoning about Logistic Regression with Missing Features
What to Expect of Classifiers? Reasoning about Logistic Regression with Missing Features
复制标题
DOI：
10.24963/ijcai.2019/377
复制
发表时间：
2019-03
期刊：
影响因子：
--
通讯作者：
Pasha Khosravi;Yitao Liang;YooJung Choi;Guy Van den Broeck
中科院分区：
文献类型：
--
作者：
Pasha Khosravi;Yitao Liang;YooJung Choi;Guy Van den Broeck
尽管判别性分类器通常会产生强大的预测性能，但预测时间的缺少特征值仍然可能是一个挑战。分类器在某些替代缺失值的方式下的行为可能不会如预期的那样，因为它们固有地对培训的数据分布做出了假设。在本文中，我们提出了一个新颖的框架，该框架通过计算针对特征分布的预期预测来对丢失特征进行分类。此外，我们使用几何编程来学习嵌入给定的逻辑回归分类器的天真贝叶斯分布，并可以有效地进行预期的预测。经验评估表明，我们的模型具有与观察到的所有功能的逻辑回归相同的性能，并且在预测期间缺少特征时，超出标准插补技术的表现。此外，我们证明我们的方法可用于通过删除不影响分类的特征来生成逻辑回归分类的``足够解释''。
While discriminative classifiers often yield strong predictive performance, missing feature values at prediction time can still be a challenge. Classifiers may not behave as expected under certain ways of substituting the missing values, since they inherently make assumptions about the data distribution they were trained on. In this paper, we propose a novel framework that classifies examples with missing features by computing the expected prediction with respect to a feature distribution. Moreover, we use geometric programming to learn a naive Bayes distribution that embeds a given logistic regression classifier and can efficiently take its expected predictions. Empirical evaluations show that our model achieves the same performance as the logistic regression with all features observed, and outperforms standard imputation techniques when features go missing during prediction time. Furthermore, we demonstrate that our method can be used to generate ``sufficient explanations'' of logistic regression classifications, by removing features that do not affect the classification.