Sklearn - 朴素贝叶斯

---

官方相关文档：Sklearn - 1.9. Naive Bayes
https://scikit-learn.org/stable/modules/naive_bayes.html

---

载入数据

from sklearn import datasets 

iris = datasets.load_iris()
iris_feature = iris.data 
iris_target = iris.target 

---

1、为连续数据 训练分类器

from sklearn.naive_bayes import GaussianNB 

# 创建高斯朴素贝叶斯对象
clf = GaussianNB() 

model = clf.fit(iris_feature, iris_target) 

new_ob = [[4, 4, 4, 0.4]]
model.predict(new_ob)  #  array([1]) 

# 给定每个分类 先验概率，创建一个高斯朴素贝叶斯对象
clf = GaussianNB(priors=[0.25, 0.25, 0.5])
model = clf.fit(iris_feature, iris_target) 

model.predict(new_ob) # array([1])

---

2、为离散数据和计数数据 训练分类器

from sklearn.naive_bayes import MultinomialNB
from sklearn.feature_extraction.text import CountVectorizer  
import numpy as np 


text_data = np.array([ 
    'I love Brazil.',
    'Brazil is better',
    'France beats both', 
])


# 创建词袋
count = CountVectorizer()
bag_of_words = count.fit_transform(text_data)

# 创建特征矩阵
features = bag_of_words.toarray() 
target = np.array([0, 0, 1]) 

# 给定每个分类的先验概率，创建一个多项式朴素贝叶斯对象
clf = MultinomialNB(class_prior=[0.25, 0.25]) 

# 训练模型
model = clf.fit(features, target) 

new_ob = [[0, 0, 0, 1, 0, 1, 0]] 

model.predict(new_ob)  # array([0])

---

3、为具有二元特征的数据，训练朴素贝叶斯分类器

from sklearn.naive_bayes import BernoulliNB 

features = np.random.randint(2, size=(100, 3))

# 创建二元目标向量
target = np.random.randint(2, size=(100, 1)).ravel() 

target

array([1, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1,
       0, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 0, 1,
       1, 1, 0, 1, 0, 1, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0,
       0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0,
       1, 0, 0, 1, 1, 0, 0, 1, 1, 0, 1, 1])

 # 给定每个分类的先验概率，创建一个多项式的朴素贝叶斯对象
    
clf = BernoulliNB(class_prior=[0.25, 0.5]) 

model = clf.fit(features, target) 

 # 想设置统一的先验概率，可设置 fit_prior=False 
    
model_uniform_prior = BernoulliNB(class_prior=None, fit_prior=False)

---

4、校准预测概率

from sklearn.calibration import CalibratedClassifierCV

clf = GaussianNB() 

# 创建使用 sigmoid 校准 调校过的 交叉验证模型
clf_sigmoid = CalibratedClassifierCV(clf, cv=2, method='sigmoid')

# 校准概率
clf_sigmoid.fit(iris_feature, iris_target) 
#    CalibratedClassifierCV(base_estimator=GaussianNB(), cv=2)

new_ob = [[2.6, 2.6, 2.6, 0.4]]
clf_sigmoid.predict_proba(new_ob) 
#    array([[0.31859969, 0.63663466, 0.04476565]])

# 训练一个 高斯朴素贝叶斯 分类器，来预测观察值的分类概率

clf.fit(iris_feature, iris_target).predict_proba(new_ob) 
#    array([[2.31548432e-04, 9.99768128e-01, 3.23532277e-07]])

# 查看校准后的概率

clf_sigmoid.predict_proba(new_ob)
#    array([[0.31859969, 0.63663466, 0.04476565]])

---

2023-03-29