Calculate Precision & Recall & F-Measure & Accuracy with Confusion Matrix (TP, TN, FP, FN)
Confusion Matrix ::
| Positive Prediction | Negative Prediction | |
|---|---|---|
| Positive Class | True Positive (TP) | False Negative (FN) |
| Negative Class | False Positive (FP) | True Negative (TN) |
Precision quantifies Formula :
Precision = TruePositives / (TruePositives + FalsePositives)
Precision = TP / ( TP + FP )
Recall quantifies Formula :
Recall = TruePositives / (TruePositives + FalseNegatives)
Recall = TP / ( TP + FN )
F-Measure quantifies Formula :
F-Measure = (2 * Precision * Recall) / (Precision + Recall)
Accuracy quantifies Formula :
accuracy = (TruePositives + TrueNegative) / (TruePositives + FalsePositive + False Negative + TrueNegative)
accuracy = (TP + TN) / (TP + FP + FN + TN)
Code (python class) :
class confusionMatrixClassification:
# Author : Amir Shokri
# Email : amirsh.nll@gmail.com
# Date : Dec 2021
# git : https://github.com/amirshnll/confusion-matrix-classification/
def __init__(self):
self.tp = 0.0
self.tf = 0.0
self.fp = 0.0
self.fn = 0.0
def setTP(self, tp):
self.tp = float(tp)
def getTP(self):
return float(self.tp)
def setTF(self, tf):
self.tf = float(tf)
def getTF(self):
return float(self.tf)
def setFP(self, fp):
self.fp = float(fp)
def getFP(self):
return float(self.fp)
def setFN(self, fn):
self.fn = float(fn)
def getFN(self):
return float(self.fn)
def precision(self):
try:
return ( self.getTP() + self.getTF() ) / ( self.getTP() + self.getTF() + self.getFP() + self.getFN() )
except:
return -1 # error
def accuracy(self):
try:
return self.getTP() / ( self.getTP() + self.getFP() )
except:
return -1 # error
def recall(self):
try:
return self.getTP() / ( self.getTP() + self.getFN() )
except:
return -1 # error
def fmeasure(self):
precision_value = self.precision()
recall_value = self.recall()
if precision_value == -1 or recall_value == -1:
return -1 # error
try:
return ( 2 * precision_value ) / ( precision_value + recall_value )
except:
return -1 # error
Code (usage) :
cmc = confusionMatrixClassification()
cmc.setTP(90)
cmc.setFP(30)
cmc.setFN(10)
print("Accuracy : " + str(cmc.accuracy())) # 0.75
print("Precision : " + str(cmc.precision())) # 0.6923076923076923
print("Recall : " + str(cmc.recall())) # 0.9
print("F-Measure : " + str(cmc.fmeasure())) # 0.8695652173913042