classifier

1. a positive example classified as positive. This is a true positive.
2. a positive example misclassified as negative. This is a false negative.
3. a negative example classified as negative. This is a true negative.
4. a negative example misclassified as positive. This is a false positive.

 

 

 

 

Here are some metrics you’ll likely come across:

• true positive rate = TP/(TP+FN) = 1 − false negative rate
• false positive rate = FP/(FP+TN) = 1 − true negative rate
• sensitivity = true positive rate
• specificity = true negative rate
• positive predictive value = TP/(TP+FP)
• recall = TP / (TP+FN) = true positive rate
• precision = TP / (TP+FP)
• F-score is the harmonic mean of precision and recall: 
  
• G-score is the geometric mean of precision and recall: 
  

 

 

 

from: http://svds.com/post/basics-classifier-evaluation-part-1