sklearn score

from sklearn import metrics
y_true = [0, 1, 2, 0, 1, 2,1]
y_pred = [0, 2, 1, 0, 0, 1,1]
a = metrics.precision_score(y_true, y_pred, average='weighted')
b = metrics.recall_score(y_true, y_pred, average='weighted')
c = metrics.f1_score(y_true, y_pred, average='weighted')

print(a,b,c,2*a*b/(a+b))

micro 爲不分類別進行計算，

macro 分母爲類別總量，例子中爲3

sample 分母爲樣本總量，例子中爲7

weight 按照各類別的樣本量爲權重進行計算

http://scikit-learn.org/stable/modules/model_evaluation.html#multiclass-and-multilabel-classification