机器学习sklearn（83）：算法实例（40）分类（19）朴素贝叶斯（二） 不同分布下的贝叶斯（一） 高斯朴素贝叶斯GaussianNB

1 认识高斯朴素贝叶斯

 

 

 

1. 展示我所使用的设备以及各个库的版本

%%cmd
pip install watermark
#在这里必须分开cell，魔法命令必须是一个cell的第一部分内容
#注意load_ext这个命令只能够执行一次，再执行就会报错，要求用reload命令
%load_ext watermark
%watermark -a "TsaiTsai" -d -v -m -p numpy,pandas,matplotlib,scipy,sklearn

2. 导入需要的库和数据

import numpy as np
import matplotlib.pyplot as plt
from sklearn.naive_bayes import GaussianNB
from sklearn.datasets import load_digits
from sklearn.model_selection import train_test_split
digits = load_digits()
X, y = digits.data, digits.target
Xtrain,Xtest,Ytrain,Ytest = train_test_split(X,y,test_size=0.3,random_state=420)

3. 建模，探索建模结果 

gnb = GaussianNB().fit(Xtrain,Ytrain) #查看分数
acc_score = gnb.score(Xtest,Ytest)
acc_score
#查看预测结果
Y_pred = gnb.predict(Xtest) #查看预测的概率结果
prob = gnb.predict_proba(Xtest)
prob.shape
prob.shape #每一列对应一个标签下的概率
prob[1,:].sum() #每一行的和都是一
prob.sum(axis=1)

4. 使用混淆矩阵来查看贝叶斯的分类结果

from sklearn.metrics import confusion_matrix as CM
CM(Ytest,Y_pred) 
#注意，ROC曲线是不能用于多分类的。多分类状况下最佳的模型评估指标是混淆矩阵和整体的准确度

2 探索贝叶斯：高斯朴素贝叶斯擅长的数据集

import numpy as np
import matplotlib.pyplot as plt
from matplotlib.colors import ListedColormap
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.datasets import make_moons, make_circles, make_classification
from sklearn.naive_bayes import GaussianNB
h = .02
names = ["Multinomial","Gaussian","Bernoulli","Complement"]
classifiers = [MultinomialNB(),GaussianNB(),BernoulliNB(),ComplementNB()]
X, y = make_classification(n_features=2, n_redundant=0, n_informative=2,
                           random_state=1, n_clusters_per_class=1)
rng = np.random.RandomState(2) X += 2 * rng.uniform(size=X.shape)
linearly_separable = (X, y)
datasets = [make_moons(noise=0.3, random_state=0),
            make_circles(noise=0.2, factor=0.5, random_state=1),
            linearly_separable
           ]
figure = plt.figure(figsize=(6, 9))
i = 1
for ds_index, ds in enumerate(datasets):
    X, y = ds
    X = StandardScaler().fit_transform(X) 
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.4, 
random_state=42)
    x1_min, x1_max = X[:, 0].min() - .5, X[:, 0].max() + .5
    x2_min, x2_max = X[:, 1].min() - .5, X[:, 1].max() + .5
    array1,array2 = np.meshgrid(np.arange(x1_min, x1_max, 0.2),
                         np.arange(x2_min, x2_max, 0.2))
    cm = plt.cm.RdBu
    cm_bright = ListedColormap(['#FF0000', '#0000FF'])
    ax = plt.subplot(len(datasets), 2, i)
    if ds_index == 0:
        ax.set_title("Input data")
    ax.scatter(X_train[:, 0], X_train[:, 1], c=y_train, 
               cmap=cm_bright,edgecolors='k')
    ax.scatter(X_test[:, 0], X_test[:, 1], c=y_test, 
               cmap=cm_bright, alpha=0.6,edgecolors='k')
    ax.set_xlim(array1.min(), array1.max())
    ax.set_ylim(array2.min(), array2.max())
    ax.set_xticks(())
    ax.set_yticks(())
    i += 1
    ax = plt.subplot(len(datasets),2,i)
    clf = GaussianNB().fit(X_train, y_train)
    score = clf.score(X_test, y_test)
    
    Z = clf.predict_proba(np.c_[array1.ravel(),array2.ravel()])[:, 1]
    Z = Z.reshape(array1.shape)
    ax.contourf(array1, array2, Z, cmap=cm, alpha=.8)
    ax.scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap=cm_bright,
               edgecolors='k')
    ax.scatter(X_test[:, 0], X_test[:, 1], c=y_test, cmap=cm_bright,
               edgecolors='k', alpha=0.6)
    
    ax.set_xlim(array1.min(), array1.max())
    ax.set_ylim(array2.min(), array2.max())
    ax.set_xticks(())
    ax.set_yticks(())
    if ds_index == 0:
        ax.set_title("Gaussian Bayes")
   
    ax.text(array1.max() - .3, array2.min() + .3, ('{:.1f}%'.format(score*100)),
            size=15, horizontalalignment='right')
    i += 1
plt.tight_layout()
plt.show()

 

3 探索贝叶斯：高斯朴素贝叶斯的拟合效果与运算速度

1. 首先导入需要的模块和库 

import numpy as np
import matplotlib.pyplot as plt
from sklearn.naive_bayes import GaussianNB
from sklearn.svm import SVC
from sklearn.ensemble import RandomForestClassifier as RFC
from sklearn.tree import DecisionTreeClassifier as DTC
from sklearn.linear_model import LogisticRegression as LR
from sklearn.datasets import load_digits
from sklearn.model_selection import learning_curve
from sklearn.model_selection import ShuffleSplit
from time import time
import datetime

2. 定义绘制学习曲线的函数

def plot_learning_curve(estimator,title, X, y, 
                        ax, #选择子图
                        ylim=None, #设置纵坐标的取值范围
                        cv=None, #交叉验证
                        n_jobs=None #设定索要使用的线程
                       ):
    train_sizes, train_scores, test_scores = learning_curve(estimator, X, y
                                                           ,cv=cv,n_jobs=n_jobs)    
    ax.set_title(title)
    if ylim is not None:
        ax.set_ylim(*ylim)
    ax.set_xlabel("Training examples")
    ax.set_ylabel("Score")
    ax.grid() #显示网格作为背景，不是必须
    ax.plot(train_sizes, np.mean(train_scores, axis=1), 'o-'
           , color="r",label="Training score")
    ax.plot(train_sizes, np.mean(test_scores, axis=1), 'o-'
           , color="g",label="Test score")
    ax.legend(loc="best")
    return ax

3. 导入数据，定义循环

digits = load_digits()
X, y = digits.data, digits.target
X.shape
X #是一个稀疏矩阵
title = ["Naive Bayes","DecisionTree","SVM, RBF kernel","RandomForest","Logistic"] 
model = [GaussianNB(),DTC(),SVC(gamma=0.001)
         ,RFC(n_estimators=50),LR(C=.1,solver="lbfgs")]
cv = ShuffleSplit(n_splits=50, test_size=0.2, random_state=0)

4. 进入循环，绘制学习曲线

fig, axes = plt.subplots(1,5,figsize=(30,6))
for ind,title_,estimator in zip(range(len(title)),title,model):
    times = time()
    plot_learning_curve(estimator, title_, X, y,
                        ax=axes[ind], ylim = [0.7, 1.05],n_jobs=4, cv=cv)
    print("{}:{}".format(title_,datetime.datetime.fromtimestamp(time()-
times).strftime("%M:%S:%f")))
plt.show()