def accuracyMicro(y_test, predictions):
"""
Accuracy Micro of our model
Params
======
y_test : sparse or dense matrix (n_samples, n_labels)
Matrix of labels used in the test phase
predictions: sparse or dense matrix (n_samples, n_labels)
Matrix of predicted labels given by our model
Returns
=======
accuracymicro : float
Accuracy Micro of our model
"""
accuracymicro = 0.0
TP, FP, TN, FN = multilabelConfussionMatrix(y_test, predictions)
TPMicro, FPMicro, TNMicro, FNMicro = multilabelMicroConfussionMatrix(
TP, FP, TN, FN)
if (TPMicro + FPMicro + TNMicro + FNMicro) != 0:
accuracymicro = float(
(TPMicro + TNMicro) / (TPMicro + FPMicro + TNMicro + FNMicro))
return accuracymicro
def fbetaMicro(y_test, predictions, beta=1):
"""
FBeta Micro of our model
Params
======
y_test : sparse or dense matrix (n_samples, n_labels)
Matrix of labels used in the test phase
predictions: sparse or dense matrix (n_samples, n_labels)
Matrix of predicted labels given by our model
Returns
=======
fbetamicro : float
FBeta Micro of our model
"""
fbetamicro = 0.0
TP, FP, TN, FN = multilabelConfussionMatrix(y_test, predictions)
TPMicro, FPMicro, TNMicro, FNMicro = multilabelMicroConfussionMatrix(
TP, FP, TN, FN)
num = float((1 + pow(beta, 2)) * TPMicro)
den = float((1 + pow(beta, 2)) * TPMicro + pow(beta, 2) * FNMicro +
FPMicro)
fbetamicro = float(num / den)
return fbetamicro
def precisionMicro(y_test, predictions):
"""
Precision Micro of our model
Params
======
y_test : sparse or dense matrix (n_samples, n_labels)
Matrix of labels used in the test phase
predictions: sparse or dense matrix (n_samples, n_labels)
Matrix of predicted labels given by our model
Returns
=======
precisionmicro : float
Precision micro of our model
"""
precisionmicro = 0.0
TP, FP, TN, FN = multilabelConfussionMatrix(y_test, predictions)
TPMicro, FPMicro, TNMicro, FNMicro = multilabelMicroConfussionMatrix(
TP, FP, TN, FN)
if (TPMicro + FPMicro) != 0:
precisionmicro = float(TPMicro / (TPMicro + FPMicro))
return precisionmicro
