import numpy as np
from sklearn.metrics import accuracy_score
from sklearn.metrics import confusion_matrix


y_true = ['A']*100 + ['B']*100 + ['C']*100
y_pred = ['A']*30 + ['B']*50 + ['C']*20 + ['A']*20 + ['B']*60 + ['C']*20 + ['A']*10 + ['B']*10 + ['C']*80

accuracy_score(y_true, y_pred)

bw@xps13 ~ $ def one_class_accuracy(actual, pred):
............     class_count = len(set(actual))
............     cm = confusion_matrix(actual, pred)
............     cm_total = cm
............
............     TP_sum, FP_sum, TN_sum, FN_sum = 0, 0, 0, 0
............     ACC_sum, SEN_sum, SPE_sum, ERR_sum = 0, 0, 0, 0
............     for c in range(class_count):
............         TP, FP, TN, FN = 0, 0, 0, 0
............         for i in range(class_count):
............             for j in range(class_count):
............                 if i == c and j == c:
............                     TP += cm_total[i, j]
............                 elif i == c and j != c:
............                     FN += cm_total[i, j]
............                 elif i != c and j == c:
............                     FP += cm_total[i, j]
............                 else:
............                     TN += cm_total[i, j]
............                     # print(cm_total[i, j])
............         TP_sum += TP
............         FP_sum += FP
............         TN_sum += TN
............         FN_sum += FN
............
............         ACC = (TP + TN) / (TP + TN + FP + FN) * 100
............         SEN = TP / (TP + FN) * 100
............         SPE = TN / (TN + FP) * 100
............         ACC_sum += ACC
............         SEN_sum += SEN
............         SPE_sum += SPE
............
............     ACC_sum /= class_count
............     SEN_sum /= class_count
............     SPE_sum /= class_count
............
............     results = {
............         'TP': int(TP_sum),
............         'TN': int(TN_sum),
............         'FP': int(FP_sum),
............         'FN': int(FN_sum),
............         'ACC': float(ACC_sum),
............         'SEN': float(SEN_sum),
............         'SPE': float(SPE_sum),
............         'CM': cm_total,
............     }
............     return results['ACC']

one_class_accuracy(y_true, y_pred)