Question

我正在尝试根据“实际”和“预测”列创建一个新列（“ Conf_Type”）。实际和预测列的输出有四种组合：（0,0），（1,1），（0,1），（1,0），我正在尝试将它们归类到新列中。

      Actual   Predicted     A         B
0         1          1  0.002753  0.997247
1         0          0  0.909696  0.090304
2         1          1  0.100924  0.899076
3         0          1  0.114059  0.885941
4         1          0  0.237289  0.762711
5         1          1  0.077710  0.922290
6         0          0  0.677748  0.322252
7         1          1  0.096327  0.903673
8         0          1  0.039741  0.960259
9         0          1  0.096884  0.903116
10        1          1  0.045345  0.954655

我尝试使用“ for”循环，但始终会出现值错误。

Conf_Type = []

for row in visual:
    if visual['Actual'] == 1 & visual['Predicted'] == 1:
        Conf_Type.append('True Negative')
    elif visual['Actual'] == 0 & visual['Predicted'] == 0:
        Conf_Type.append('True Positive')
    elif visual['Actual'] == 1 & visual['Predicted'] == 0:
        Conf_Type.append('False Positive')
    elif visual['Actual'] == 0 & visual['Predicted'] == 1:
        Conf_Type.append('False Negative')

visual['Conf_Type'] = Conf_Type

ValueError: The truth value of a Series is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

我想知道在这种情况下“ for”循环是正确的还是应该使用其他方法。

Answer 1

您应该像在这里一样使用apply。

number_meaning = ['true positive', 'false negative', 'false positive', 'true negative']
df['categorized'] = df['Actual'] + 2 * df['Predicted']
df['result'] = df['categorized'].apply(lambda x: number_meaning[x])

或使用您定义的函数：

def meaning(row):
    if row['Actual'] == 1 & row['Predicted'] == 1:
        return 'True Negative'
    elif row['Actual'] == 0 & row['Predicted'] == 0:
        return 'True Positive'
    elif row['Actual'] == 1 & row['Predicted'] == 0:
        return 'False Positive'
    elif row['Actual'] == 0 & row['Predicted'] == 1:
        return 'False Negative'
df['result'] = df.apply(meaning, axis=1)

Answer 2

# iterate every row by row number
for i in list(df.index):
    if df.loc[i,"Actual"] ==df.loc[i,"Predicted"]==0 or df.loc[i,"Actual"] ==df.loc[i,"Predicted"]==1:
        # set up the value of row 'i', column 'Conf_Type' based on this condition
        df.loc[i, "Conf_Type"] = True
    else:
        df.loc[i, "Conf_Type"] = False

基于两个过滤列的循环

2 个答案: