Terms

Terms and abbreviations used in the description of Binary classification assessment visualizer are listed here.

• Сondition positive (P): the real number of positive cases.
• Сondition negative (N): the real number of negative cases.
• True positive (TP): the number of correctly classified positive cases.
• True negative (TN): the number of correctly classified negative cases.
• False positive (FP): the number of incorrectly classified positive cases. It is type I error.
• False negative (FN): the number of incorrectly classified negative cases. It is type II error.
• True positive rate (TPR): rate of correctly classified positive cases based on the total number of positive cases. TPR is called sensitivity of classification.
  $TPR=\frac{TP}{P}=\frac{TP}{TP+FN}=1-FNR$.
• True negative rate (TNR): rate of correctly classified negative cases based on the total number of negative cases. TNR is called Specificity of classification.
  $TNR=\frac{TN}{N}=\frac{TN}{TN+FP}=1-FPR$.
• Positive predictive value (PPV): rate of correctly classified positive cases based on the total number of the cases classified as the positive ones. PPV is called Positive predictive value of classification.
  $PPV=\frac{TP}{TP+FP}$.
• Negative predictive value (NPV): rate of correctly classified negative cases based on the total number of the cases classified as the negative ones. PPV is called Negative prediction value of classification.
  $NPV=\frac{TN}{TN+FN}$.
• False negative rate (FNR): rate of incorrectly classified negative cases based on the total number of positive cases. For FNR 1 - Sensitivity. term is frequently used $FNR=\frac{FN}{P}=\frac{FN}{FN+TP}=1-TPR$.
• False positive rate (FPR): rate of incorrectly classified positive cases based on the total number of negative cases. For FPR 1 - Specificity. term is frequently used $FPR=\frac{FP}{N}=\frac{FP}{FP+TN}=1-TNR$.
• Overall classification rate (OCR): rate of correctly classified cases based on the total number of cases. OCR is called Overall classification rate.
  $OCR=\frac{TP+TN}{P+N}=\frac{TP+TN}{TP+TN+FP+FN}$.
• F₁ score means the average harmonic value of Accuracy and Sensitivity. The metric is also called F measure.
  $F_1=2\cdot \frac{PPV\cdot TPR}{PPV+TPR}=\frac{2TP}{2TP+FP+FN}$.
• Matthews correlation coefficient (MCC) — Matthews correlation coefficient.
  $MCC=\frac{TP\cdot TN-FP\cdot FN}{\sqrt{(TP+FP)(TP+FN)(TN+FP)(TN+FN)}}$.
• Overall precision rate (OPR) means the arithmetical mean of Positive predictive value and Negative predictive value. OPR means the Overall Precision Rate.
  $OPR=\frac{PPV+NPV}{2}$.