import pandas as pd
import numpy as np
from sklearn import metrics
from sklearn.metrics import confusion_matrix, classification_report, average_precision_score, precision_recall_curve, accuracy_score, confusion_matrix, average_precision_score
def episode_metrics(model, x, parameters, rhythm, out_message=False):
""" arg
"""
# Rhythmn predictions
predictions_qa, predictions_r = model.predict(x)
# If the estimated prob of AF(column 1) is higher than the estimated prob of non-AF(column 0)
# than consider the window as an AF window.
y_predictions = np.argmax(predictions_r, axis=1)
y_truth = np.argmax(rhythm, axis=1)
# Confusion_matrix
cf = confusion_matrix(y_truth, y_predictions, labels=[0, 1])
TN, FP, FN, TP = cf.ravel()
support = TN+FP+FN+TP
# Sensitivity, recall, or true positive rate
TPR = TP/(TP+FN)
# Specificity or true negative rate
TNR = TN/(TN+FP)
# Precision or positive predictive value
PPV = TP/(TP+FP)
# Negative predictive value
NPV = TN/(TN+FN)
# Fall out or false positive rate
FPR = FP/(FP+TN)
# False negative rate
FNR = FN/(TP+FN)
# F1 score
f1 = metrics.f1_score(y_truth, y_predictions, average=None)
# selecting f1score for positive case if present
f1_pos = f1[1] if len(f1) > 1 else f1[0]
# Area under precision recall curve
#auprc = metrics.average_precision_score(y_truth, rhythm[:,1])
if out_message:
print(pd.DataFrame(cf).rename(columns={0: "Predicted Non-AF", 1:" Predicted AF"}, index={0: "True Non-AF", 1:"True AF"}))
print("Sensitivity/Recall: %0.4f" % TPR)
print("Specificity: %0.4f" % TNR)
print("Precision/PPV: %0.4f" % PPV)
print("Negative predictive value/NPV: %0.2f" % NPV)
print("False positive rate: %0.4f" % FPR)
print("False negative rate: %0.4f" % FNR)
print("F1 score: %0.4f" % f1_pos)
print('support: ', support)
print('\n\n\n')
episode_metrics = [TPR, TNR, PPV, NPV, FPR, FNR, f1_pos, support]
return episode_metrics
def episode_metrics_singletask(model, x, parameters, rhythm, out_message=False):
""" arg
"""
# Rhythmn predictions
predictions_r = model.predict(x)
# If the estimated prob of AF(column 1) is higher than the estimated prob of non-AF(column 0)
# than consider the window as an AF window.
y_predictions = np.argmax(predictions_r, axis=1)
y_truth = np.argmax(rhythm, axis=1)
# Confusion_matrix
cf = confusion_matrix(y_truth, y_predictions, labels=[0, 1])
TN, FP, FN, TP = cf.ravel()
support = TN+FP+FN+TP
# Sensitivity, recall, or true positive rate
TPR = TP/(TP+FN)
# Specificity or true negative rate
TNR = TN/(TN+FP)
# Precision or positive predictive value
PPV = TP/(TP+FP)
# Negative predictive value
NPV = TN/(TN+FN)
# Fall out or false positive rate
FPR = FP/(FP+TN)
# False negative rate
FNR = FN/(TP+FN)
# F1 score
f1 = metrics.f1_score(y_truth, y_predictions, average=None)
# selecting f1score for positive case if present
f1_pos = f1[1] if len(f1) > 1 else f1[0]
# Area under precision recall curve
#auprc = metrics.average_precision_score(y_truth, rhythm[:,1])
if out_message:
print(pd.DataFrame(cf).rename(columns={0: "Predicted Non-AF", 1:" Predicted AF"}, index={0: "True Non-AF", 1:"True AF"}))
print("Sensitivity/Recall: %0.4f" % TPR)
print("Specificity: %0.4f" % TNR)
print("Precision/PPV: %0.4f" % PPV)
print("Negative predictive value/NPV: %0.2f" % NPV)
print("False positive rate: %0.4f" % FPR)
print("False negative rate: %0.4f" % FNR)
print("F1 score: %0.4f" % f1_pos)
print('support: ', support)
print('\n\n\n')
episode_metrics = [TPR, TNR, PPV, NPV, FPR, FNR, f1_pos, support]
return episode_metrics
def collecting_individual_metrics(model, x, parameters, rhythm, out_message=False):
"""
"""
individual_metrics = {}
for i in np.unique(parameters['ID']):
# Sub-selecting individuals
p_indx = np.where(parameters['ID'] == i)[0]
x_pID = x[p_indx]
parameters_pID = parameters.iloc[p_indx]
rhythm_pID = rhythm[p_indx]
# Rhythmn predictions
predictions_qa, predictions_r = model.predict(x_pID)
# If the estimated prob of AF(column 1) is higher than the estimated prob of non-AF(column 0)
# than consider the window as an AF window.
y_predictions = np.argmax(predictions_r, axis=1)
y_truth = np.argmax(rhythm_pID, axis=1)
# Confusion_matrix
cf = confusion_matrix(y_truth, y_predictions, labels=[0, 1])
TN, FP, FN, TP = cf.ravel()
support = TN+FP+FN+TP
# Sensitivity, recall, or true positive rate
TPR = TP/(TP+FN)
# Specificity or true negative rate
TNR = TN/(TN+FP)
# Fall out or false positive rate
FPR = FP/(FP+TN)
# False negative rate
FNR = FN/(TP+FN)
# F1 score
f1 = metrics.f1_score(y_truth, y_predictions, average=None)
# selecting f1score for positive case if present
f1_pos = f1[1] if len(f1) > 1 else f1[0]
# Area under precision recall curve
#auprc = metrics.average_precision_score(y_test_truth, rhythm_pID[:,1])
if (TP + FN) > 0:
individual_metrics[i] = [TPR, np.NaN, np.NaN, FNR, f1_pos, support]
if (FP + TN) > 0:
individual_metrics[i] = [np.NaN, TNR, FPR, np.NaN, np.NaN, support]
if out_message:
print("PATIENT :", i , '\n')
print(pd.DataFrame(cf).rename(columns={0: "Predicted Non-AF", 1:" Predicted AF"}, index={0: "True Non-AF", 1:"True AF"}))
if (TP + FN) > 0:
print("Sensitivity/Recall: %0.4f" % TPR)
print("False negative rate: %0.4f" % FNR)
print("F1 score: %0.4f" % f1_pos)
print('support: ' , support)
if (FP + TN) > 0:
print("Specificity: %0.4f" % TNR)
print("False positive rate: %0.4f" % FPR)
print('support: ' , support)
return individual_metrics
def collecting_individual_metrics_singletask(model, x, parameters, rhythm, out_message=False):
"""
"""
individual_metrics = {}
for i in np.unique(parameters['ID']):
# Sub-selecting individuals
p_indx = np.where(parameters['ID'] == i)[0]
x_pID = x[p_indx]
parameters_pID = parameters.iloc[p_indx]
rhythm_pID = rhythm[p_indx]
# Rhythmn predictions
predictions_r = model.predict(x_pID)
# If the estimated prob of AF(column 1) is higher than the estimated prob of non-AF(column 0)
# than consider the window as an AF window.
y_predictions = np.argmax(predictions_r, axis=1)
y_truth = np.argmax(rhythm_pID, axis=1)
# Confusion_matrix
cf = confusion_matrix(y_truth, y_predictions, labels=[0, 1])
TN, FP, FN, TP = cf.ravel()
support = TN+FP+FN+TP
# Sensitivity, recall, or true positive rate
TPR = TP/(TP+FN)
# Specificity or true negative rate
TNR = TN/(TN+FP)
# Fall out or false positive rate
FPR = FP/(FP+TN)
# False negative rate
FNR = FN/(TP+FN)
# F1 score
f1 = metrics.f1_score(y_truth, y_predictions, average=None)
# selecting f1score for positive case if present
f1_pos = f1[1] if len(f1) > 1 else f1[0]
# Area under precision recall curve
#auprc = metrics.average_precision_score(y_test_truth, rhythm_pID[:,1])
if (TP + FN) > 0:
individual_metrics[i] = [TPR, np.NaN, np.NaN, FNR, f1_pos, support]
if (FP + TN) > 0:
individual_metrics[i] = [np.NaN, TNR, FPR, np.NaN, np.NaN, support]
if out_message:
print("PATIENT :", i , '\n')
print(pd.DataFrame(cf).rename(columns={0: "Predicted Non-AF", 1:" Predicted AF"}, index={0: "True Non-AF", 1:"True AF"}))
if (TP + FN) > 0:
print("Sensitivity/Recall: %0.4f" % TPR)
print("False negative rate: %0.4f" % FNR)
print("F1 score: %0.4f" % f1_pos)
print('support: ' , support)
if (FP + TN) > 0:
print("Specificity: %0.4f" % TNR)
print("False positive rate: %0.4f" % FPR)
print('support: ' , support)
return individual_metrics
Datasets
Models
