Know every term of Confusion matrix | Let's make it easy to understand !

It is the mark of a truly intelligent person to be moved by Statistics by George Bernard Shaw, Irish playwright and political activist

Every term of Confusion matrix

Let's understand through a confusion matrix of most famous dataset: Loan defaulter. Through this confusion matrix, we will try to calculate every possible metric from it.

Confusion matrix

Non defaulter means customer is paying loan timely
Defaulter means customer is not paying timely to the bank, having dues.

True Positive (TP): We predicted 20 cases will be defaulter and turns out to be true.
True Negative (TN): We predicted 121 cases are non defaulter  and it turns out to be true.
False Positive (FP): We predicted 22 cases are defaulter but turns out to be false.
False Negative (FN): We predicted 7 cases are non defaulter but turns out to be false.

Accuracy: How often is the classifier correct.
(TP+TN)/(TP+TN+FP+FN) = 82.94%

Sensitivity/Recall/True Positive rate: When it's actually Defaulter, how often does it predict yes.
TP/(TP+FN) = 74.07%

Precision: The precision metric shows the accuracy of the positive class. It measures how likely the prediction of the positive class is correct.
TP/(TP+FP) = 47.61%

Specificity/ True Negative rate: When it's actually non defaulter, how often does it predict the same.
TN/(TN+FP) = 84.61%

F-measure: It is the representation of both Precision and Recall metric. It uses harmonic mean and it's value always close to the smaller value of either precision or recall.
(2*Recall*Precision)/(Recall+Precision) = 57.96%



Feel free to comment your opinion💓