一、
努利贝叶斯类BernoulliN假设数据服从多元伯努利分布,并在此基础上应用朴素贝叶斯的训练和分类过程。多元伯努利分布简单来说,就是数据集中可以存在多个特征,但每个特征都是二分类的,可以以布尔变量表示,也可以表示为{0,1}或者{-1,1}等任意二分类组合。因此,这个类要求将样本转换为二分类特征向量,如果数据本身不是二分类的,那可以使用类中专门用来二值化的参数binarize来改变数据。
- binarize or not
import numpy as np
from sklearn.datasets import make_blobs
from sklearn.preprocessing import MinMaxScaler
from sklearn.model_selection import train_test_split
from sklearn.naive_bayes import BernoulliNB # 伯努利贝叶斯: 专门用于处理(多元)二项分布
from sklearn.metrics import brier_score_loss # 布里尔分数
class_1_samples = 500
class_2_samples = 500
centers = [[0.0, 0.0], [2.0, 2.0]] # (1000,2) ==> centers = [[0.0, 0.0, 0.0], [2.0, 2.0, 2.0]] (1000, 3)
cluster_std = [0.5, 0.5]
X, y = make_blobs(n_samples=[class_1_samples, class_2_samples], centers=centers, cluster_std=cluster_std, random_state=42, shuffle=False)
print(X.shape) # [1000, 2]
print(np.unique) # array([0, 1])
Xtrain, Xtest, ytrain, ytest = train_test_split(X, y, test_size=0.3, random_state=42)
mms = MinMaxScaler().fit(Xtrain)
Xtrain = mms.transform(Xtrain)
Xtest = mms.transform(Xtest)
bnl_ = BernoulliNB()
bnl_.fit(Xtrain, ytrain) # 不设置二值化
print(bnl_.score(Xtest, ytest))
bnl = BernoulliNB(binarize=0.5)
bnl.fit(Xtrain, ytrain) # 设置二值化阈值为0.5,将特征二值化
print(bnl.score(Xtest, ytest))
brier_score_loss(ytest, bnl.predict_proba(Xtest)[:, 1], pos_label=1)