R混淆矩阵敏感性和特异性标记 [英] R Confusion Matrix sensitivity and specificity labeling
问题描述
我使用 R v3.3.2 和 Caret 6.0.71(即最新版本)来构建逻辑回归分类器.我正在使用混淆矩阵函数来创建用于判断其性能的统计数据.
I am using R v3.3.2 and Caret 6.0.71 (i.e. latest versions) to construct a logistic regression classifier. I am using the confusionMatrix function to create stats for judging its performance.
logRegConfMat <-混淆矩阵(logRegPrediction, valData[,"Seen"])
logRegConfMat <- confusionMatrix(logRegPrediction, valData[,"Seen"])
- 参考 0,预测 0 = 30
- 参考文献 1,预测 0 = 14
- 参考 0,预测 1 = 60
- 参考文献 1,预测 1 = 164
准确度:0.7239
灵敏度:0.3333
特异性:0.9213
Accuracy : 0.7239
Sensitivity : 0.3333
Specificity : 0.9213
我的数据 (Seen) 中的目标值使用 1 表示真,0 表示假.我假设混淆矩阵中的参考(基本事实)列和谓词(分类器)行遵循相同的约定.因此我的结果显示:
The target value in my data (Seen) uses 1 for true and 0 for false. I assume the Reference (Ground truth) columns and Predication (Classifier) rows in the confusion matrix follow the same convention. Therefore my results show:
- 真阴性 (TN) 30
- 真阳性 (TP) 164
- 假阴性 (FN) 14
- 误报 (FP) 60
问题:为什么灵敏度为 0.3333,特异性为 0.9213?我会认为这是相反的 - 见下文.
我不愿意相信 R 混淆矩阵函数中存在错误,因为没有报告任何内容,这似乎是一个重大错误.
I am reluctant to believe that there is bug in the R confusionMatrix function as nothing has been reported and this seems to be a significant error.
大多数关于计算特异性和敏感性的参考文献将它们定义如下 - 即 www.medcalc.org/calc/diagnostic_test.php
Most references about calculating specificity and sensitivity define them as follows - i.e. www.medcalc.org/calc/diagnostic_test.php
- 灵敏度 = TP/(TP+FN) = 164/(164+14) = 0.9213
- 特异性 = TN/(FP+TN) = 30/(60+30) = 0.3333
推荐答案
根据文档?confusionMatrix
:
"如果只有两个因子水平,则第一个水平将用作阳性"结果."
"If there are only two factor levels, the first level will be used as the "positive" result."
因此,在您的示例中,正面结果将是 0
,并且评估指标将是错误的.要覆盖默认行为,您可以将参数 positive =
设置为正确的值,唉:
Hence in your example positive result will be 0
, and evaluation metrics will be the wrong way around. To override default behaviour, you can set the argument positive =
to the correct value, alas:
confusionMatrix(logRegPrediction, valData[,"Seen"], positive = "1")
这篇关于R混淆矩阵敏感性和特异性标记的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!