我正在尝试在我的文章上执行留一个CV,但是当我运行该程序时,我得到100%的准确性,我无法弄清楚我错过了什么。这是我的代码:
import sklearn
from sklearn.datasets import load_files
import numpy as np
from sklearn.cross_validation import cross_val_score, LeaveOneOut
from scipy.stats import sem
from sklearn.naive_bayes import MultinomialNB
bunch = load_files('corpus', shuffle = False)
X = bunch.data
y = bunch.target
from sklearn.feature_extraction.text import CountVectorizer
count_vect = CountVectorizer(stop_words = 'english')
X_counts = count_vect.fit_transform(X)
from sklearn.feature_extraction.text import TfidfTransformer
tfidf_transformer = TfidfTransformer()
X_tfidf = tfidf_transformer.fit_transform(X_counts)
estimator = MultinomialNB().fit(X_tfidf, y)
cv = LeaveOneOut(26)
scores = cross_val_score(estimator, X_tfidf, y, cv = cv)
print scores
print ("Mean score: {0:.3f} (+/-{1:.3f})").format(np.mean(scores), sem(scores))
我得到的输入数据分类相同,这有点奇怪。我的结果:
[ 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1. 1.]
Mean score: 0.577 (+/-0.099)
我的输入数据分类:
([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1])
我不明白我的LOO CV失败的地方。 :S
帮助将不胜感激。
答案 0 :(得分:1)
从最后一行打印时,您的准确度得分不是LOOCV 0.577吗?
cross_val_score函数返回每个CV折叠的分数数组(默认精度)。您打印的数组scores
是准确度分数而不是预测值。