Мобильная версияThis code provides implementation of common metrics for assessing a binary classifier's hard decisions against true binary labels, including:
- accuracy
- true positive rate and true negative rate (TPR and TNR)
- positive predictive value and negative predictive value (PPV and NPV)
....
import numpy as np def calc_TP_TN_FP_FN(ytrue_N, yhat_N): ''' Count the four possible states of true and predicted binary values. Args ---- ytrue_N : 1D array, shape (n_examples,) = (N,) All values must be either 0 or 1. Will be cast to int dtype. Each entry represents the binary 'true' label of one example One entry per example in current dataset yhat_N : 1D array, shape (n_examples,) = (N,) All values must be either 0 or 1. Will be cast to int dtype. Each entry represents a predicted label for one example One entry per example in current dataset. Needs to be same size as ytrue_N. Returns ------- TP : int Number of true positives TN : int Number of true negatives FP : int Number of false positives FN : int Number of false negatives Examples -------- >>> N = 8 >>> ytrue_N = np.asarray([0., 0., 0., 0., 1., 1., 1., 1.]) >>> yhat_N = np.asarray([0., 0., 1., 0., 1., 1., 0., 0.]) >>> TP, TN, FP, FN = calc_TP_TN_FP_FN(ytrue_N, yhat_N) >>> TP 2 >>> TN 3 >>> FP 1 >>> FN 2 >>> np.allclose(TP + TN + FP + FN, N) True ''' # Cast input to integer just to be sure we're getting what's expected ytrue_N = np.asarray(ytrue_N, dtype=np.int32) yhat_N = np.asarray(yhat_N, dtype=np.int32) # TODO fix by calculating the number of true pos, true neg, etc. TP = 0 TN = 0 FP = 0 FN = 0 return TP, TN, FP, FN def calc_ACC(ytrue_N, yhat_N): ''' Compute the accuracy of provided predicted binary values. Args ---- ytrue_N : 1D array of floats, shape (n_examples,) = (N,) All values must be either 0.0 or 1.0. Each entry represents the binary 'true' label of one example One entry per example in current dataset yhat_N : 1D array of floats, shape (n_examples,) = (N,) All values must be either 0.0 or 1.0. Each entry represents a predicted label for one example One entry per example in current dataset. Needs to be same size as ytrue_N. Returns ------- acc : float Accuracy = ratio of number correct over total number of examples Examples -------- >>> N = 8 >>> ytrue_N = np.asarray([0., 0., 0., 0., 1., 1., 1., 1.]) >>> yhat_N = np.asarray([0., 0., 1., 0., 1., 1., 0., 0.]) >>> acc = calc_ACC(ytrue_N, yhat_N) >>> print("%.3f" % acc) 0.625 ''' # TODO compute accuracy # You should *use* your calc_TP_TN_FP_FN function from above # Hint: make sure denominator will never be exactly zero # by adding a small value like 1e-10 return 0.0 def calc_TPR(ytrue_N, yhat_N): ''' Compute the true positive rate of provided predicted binary values. Also known as the recall. Args ---- ytrue_N : 1D array of floats, shape (n_examples,) = (N,) All values must be either 0.0 or 1.0. Each entry represents the binary 'true' label of one example One entry per example in current dataset yhat_N : 1D array of floats, shape (n_examples,) = (N,) All values must be either 0.0 or 1.0. Each entry represents a predicted label for one example One entry per example in current dataset. Needs to be same size as ytrue_N. Returns ------- tpr : float TPR = ratio of true positives over total labeled positive Examples -------- >>> N = 8 >>> ytrue_N = np.asarray([0., 0., 0., 0., 1., 1., 1., 1.]) >>> yhat_N = np.asarray([0., 0., 1., 0., 1., 1., 0., 0.]) >>> tpr = calc_TPR(ytrue_N, yhat_N) >>> print("%.3f" % tpr) 0.500 # Verify what happens with empty input >>> empty_val = calc_TPR([], []) >>> print("%.3f" % empty_val) 0.000 ''' # TODO compute TPR # You should *use* your calc_TP_TN_FP_FN function from above # Hint: make sure denominator will never be exactly zero # by adding a small value like 1e-10 return 0.0 def calc_TNR(ytrue_N, yhat_N): ''' Compute the true negative rate of provided predicted binary values. Args ---- ytrue_N : 1D array of floats, shape (n_examples,) = (N,) All values must be either 0.0 or 1.0. Each entry represents the binary 'true' label of one example One entry per example in current dataset yhat_N : 1D array of floats, shape (n_examples,) = (N,) All values must be either 0.0 or 1.0. Each entry represents a predicted label for one example One entry per example in current dataset. Needs to be same size as ytrue_N. Returns ------- tnr : float TNR = ratio of true negatives over total labeled negative. Examples -------- >>> N = 8 >>> ytrue_N = np.asarray([0., 0., 0., 0., 1., 1., 1., 1.]) >>> yhat_N = np.asarray([0., 0., 1., 0., 1., 1., 0., 0.]) >>> tnr = calc_TNR(ytrue_N, yhat_N) >>> print("%.3f" % tnr) 0.750 # Verify what happens with empty input >>> empty_val = calc_TNR([], []) >>> print("%.3f" % empty_val) 0.000 ''' # TODO compute TNR # You should *use* your calc_TP_TN_FP_FN function from above # Hint: make sure denominator will never be exactly zero # by adding a small value like 1e-10 return 0.0 def calc_PPV(ytrue_N, yhat_N): ''' Compute positive predictive value of provided predicted binary values. Also known as the precision. Args ---- ytrue_N : 1D array of floats, shape (n_examples,) = (N,) All values must be either 0.0 or 1.0. Each entry represents the binary 'true' label of one example One entry per example in current dataset yhat_N : 1D array of floats, shape (n_examples,) = (N,) All values must be either 0.0 or 1.0. Each entry represents a predicted label for one example One entry per example in current dataset. Needs to be same size as ytrue_N. Returns ------- ppv : float PPV = ratio of true positives over total predicted positive. Examples -------- >>> N = 8 >>> ytrue_N = np.asarray([0., 0., 0., 0., 1., 1., 1., 1.]) >>> yhat_N = np.asarray([0., 0., 1., 0., 1., 1., 0., 0.]) >>> ppv = calc_PPV(ytrue_N, yhat_N) >>> print("%.3f" % ppv) 0.667 # Verify what happens with empty input >>> empty_val = calc_PPV([], []) >>> print("%.3f" % empty_val) 0.000 ''' # TODO compute PPV # You should *use* your calc_TP_TN_FP_FN function from above # Hint: make sure denominator will never be exactly zero # by adding a small value like 1e-10 return 0.0 def calc_NPV(ytrue_N, yhat_N): ''' Compute negative predictive value of provided predicted binary values. Args ---- ytrue_N : 1D array of floats, shape (n_examples,) = (N,) All values must be either 0.0 or 1.0. Each entry represents the binary 'true' label of one example One entry per example in current dataset yhat_N : 1D array of floats, shape (n_examples,) = (N,) All values must be either 0.0 or 1.0. Each entry represents a predicted label for one example One entry per example in current dataset. Needs to be same size as ytrue_N. Returns ------- npv : float NPV = ratio of true negative over total predicted negative. Examples -------- >>> N = 8 >>> ytrue_N = np.asarray([0., 0., 0., 0., 1., 1., 1., 1.]) >>> yhat_N = np.asarray([0., 0., 1., 0., 1., 1., 0., 0.]) >>> npv = calc_NPV(ytrue_N, yhat_N) >>> print("%.3f" % npv) 0.600 # Verify what happens with empty input >>> empty_val = calc_NPV([], []) >>> print("%.3f" % empty_val) 0.000 ''' # TODO compute NPV # You should *use* your calc_TP_TN_FP_FN function from above # Hint: make sure denominator will never be exactly zero # by adding a small value like 1e-10 return 0.0 How can I fix this code?
Источник: https://stackoverflow.com/questions/780 ... redictions
