ML 监督学习 分类 朴素贝叶斯分类器

朴素贝叶斯分类

基于贝叶斯定理来计算概率和条件概率，且假设每个属性独立地对分类结果产生影响

为什么称为“朴素的”？
因为朴素贝叶斯假定所有属性都是同等重要且相互独立的。

例子

高斯贝叶斯分类

Gaussian

3种不同的贝叶斯分类

Bernoulli Multinomial Gaussian

拉普拉斯修正(Laplacian correction)

如果训练集中未出现属性值，条件概率的值为0。为了避免这种情况，在估计概率值时通常要进行"平滑"，常用'拉普拉斯修正"。

Laplacian correction

例子

代码

#GausianNB
from sklearn import datasets
from sklearn import metrics
from sklearn.naive_bayes import GaussianNB

dataset = datasets.load_iris()

model = GaussianNB()
model.fit(dataset.data, dataset.target)

expected = dataset.target
predicted = model.predict(dataset.data)

print(metrics.classification_report(expected, predicted))
print(metrics.confusion_matrix(expected, predicted))

Output

文本分类

#####Text Classification
from sklearn.datasets import fetch_20newsgroups
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB,GaussianNB
from sklearn.pipeline import make_pipeline
from sklearn.metrics import classification_report, confusion_matrix

categories = ['alt.atheism', 'talk.religion.misc',
              'comp.graphics', 'sci.space']
newsgroups_train = fetch_20newsgroups(subset='train',
                                      categories=categories)
newsgroups_test = fetch_20newsgroups(subset='test',
                                     categories=categories)

X_train = newsgroups_train.data
X_test = newsgroups_test.data

y_train = newsgroups_train.target
y_test = newsgroups_test.target

pipe = make_pipeline(TfidfVectorizer(), MultinomialNB())
pipe.fit(X_train, y_train)
y_pred = pipe.predict(X_test)

print(classification_report(y_test, y_pred))
print(confusion_matrix(y_test, y_pred))

垃圾邮件过滤
别人写的例子
https://github.com/Surya-Murali/Email-Spam-Classifier-Using-Naive-Bayes/blob/master/SpamClassifier.py

总结

Conclusion