python贝叶斯分类
常见的分类算法有：
K近邻算法
决策树
朴素贝叶斯
SVM
Logistic Regression
2. python scikit-learn 有什么算法
1，前言
很久不发文章，主要是Copy别人的总感觉有些不爽，所以整理些干货，希望相互学习吧。不啰嗦，进入主题吧，本文主要时说的为朴素贝叶斯分类算法。与逻辑回归，决策树一样，是较为广泛使用的有监督分类算法，简单且易于理解（号称十大数据挖掘算法中最简单的算法）。但其在处理文本分类，邮件分类，拼写纠错，中文分词，统计机器翻译等自然语言处理范畴较为广泛使用，或许主要得益于基于概率理论，本文主要为小编从理论理解到实践的过程记录。
2，公式推断
一些贝叶斯定理预习知识：我们知道当事件A和事件B独立时，P（AB）=P（A）（B），但如果事件不独立，则P（AB）=P（A）P（B|A）。为两件事件同时发生时的一般公式，即无论事件A和B是否独立。当然也可以写成P（AB）=P（B）P（A|B），表示若要两件事同事发生，则需要事件B发生后，事件A也要发生。
由上可知，P（A）P（B|A）= P（B）P（A|B）
推出P（B|A）=
其中P（B）为先验概率，P（B|A）为B的后验概率，P（A|B）为A的后验概率（在这里也为似然值），P（A）为A的先验概率（在这也为归一化常量）。
由上推导可知，其实朴素贝叶斯法就是在贝叶斯定理基础上，加上特征条件独立假设，对特定输入的X（样本，包含N个特征），求出后验概率最大值时的类标签Y（如是否为垃圾邮件），理解起来比逻辑回归要简单多，有木有，这也是本算法优点之一，当然运行起来由于得益于特征独立假设，运行速度也更快。
8. Python代码
# -*-coding: utf-8 -*-
importtime
fromsklearn import metrics
fromsklearn.naive_bayes import GaussianNB
fromsklearn.naive_bayes import MultinomialNB
fromsklearn.naive_bayes import BernoulliNB
fromsklearn.neighbors import KNeighborsClassifier
fromsklearn.linear_model import LogisticRegression
fromsklearn.ensemble import RandomForestClassifier
fromsklearn import tree
fromsklearn.ensemble import GradientBoostingClassifier
fromsklearn.svm import SVC
importnumpy as np
importurllib
# urlwith dataset
url ="-learning-databases/pima-indians-diabetes/pima-indians-diabetes.data"
#download the file
raw_data= urllib.request.urlopen(url)
#load the CSV file as a numpy matrix
dataset= np.loadtxt(raw_data, delimiter=",")
#separate the data from the target attributes
X =dataset[:,0:7]
#X=preprocessing.MinMaxScaler().fit_transform(x)
#print(X)
y =dataset[:,8]
print(" 调用scikit的朴素贝叶斯算法包GaussianNB ")
model= GaussianNB()
start_time= time.time()
model.fit(X,y)
print('training took %fs!' % (time.time() - start_time))
print(model)
expected= y
predicted= model.predict(X)
print(metrics.classification_report(expected,predicted))
print(metrics.confusion_matrix(expected,predicted))
print(" 调用scikit的朴素贝叶斯算法包MultinomialNB ")
model= MultinomialNB(alpha=1)
start_time= time.time()
model.fit(X,y)
print('training took %fs!' % (time.time() - start_time))
print(model)
expected= y
predicted= model.predict(X)
print(metrics.classification_report(expected,predicted))
print(metrics.confusion_matrix(expected,predicted))
print(" 调用scikit的朴素贝叶斯算法包BernoulliNB ")
model= BernoulliNB(alpha=1,binarize=0.0)
start_time= time.time()
model.fit(X,y)
print('training took %fs!' % (time.time() - start_time))
print(model)
expected= y
predicted= model.predict(X)
print(metrics.classification_report(expected,predicted))
print(metrics.confusion_matrix(expected,predicted))
print(" 调用scikit的KNeighborsClassifier ")
model= KNeighborsClassifier()
start_time= time.time()
model.fit(X,y)
print('training took %fs!' % (time.time() - start_time))
print(model)
expected= y
predicted= model.predict(X)
print(metrics.classification_report(expected,predicted))
print(metrics.confusion_matrix(expected,predicted))
print(" 调用scikit的LogisticRegression(penalty='l2')")
model= LogisticRegression(penalty='l2')
start_time= time.time()
model.fit(X,y)
print('training took %fs!' % (time.time() - start_time))
print(model)
expected= y
predicted= model.predict(X)
print(metrics.classification_report(expected,predicted))
print(metrics.confusion_matrix(expected,predicted))
print(" 调用scikit的RandomForestClassifier(n_estimators=8) ")
model= RandomForestClassifier(n_estimators=8)
start_time= time.time()
model.fit(X,y)
print('training took %fs!' % (time.time() - start_time))
print(model)
expected= y
predicted= model.predict(X)
print(metrics.classification_report(expected,predicted))
print(metrics.confusion_matrix(expected,predicted))
print(" 调用scikit的tree.DecisionTreeClassifier()")
model= tree.DecisionTreeClassifier()
start_time= time.time()
model.fit(X,y)
print('training took %fs!' % (time.time() - start_time))
print(model)
expected= y
predicted= model.predict(X)
print(metrics.classification_report(expected,predicted))
print(metrics.confusion_matrix(expected,predicted))
print(" 调用scikit的GradientBoostingClassifier(n_estimators=200) ")
model= GradientBoostingClassifier(n_estimators=200)
start_time= time.time()
model.fit(X,y)
print('training took %fs!' % (time.time() - start_time))
print(model)
expected= y
predicted= model.predict(X)
print(metrics.classification_report(expected,predicted))
print(metrics.confusion_matrix(expected,predicted))
print(" 调用scikit的SVC(kernel='rbf', probability=True) ")
model= SVC(kernel='rbf', probability=True)
start_time= time.time()
model.fit(X,y)
print('training took %fs!' % (time.time() - start_time))
print(model)
expected= y
predicted= model.predict(X)
print(metrics.classification_report(expected,predicted))
print(metrics.confusion_matrix(expected,predicted))
"""
# 预处理代码集锦
importpandas as pd
df=pd.DataFrame(dataset)
print(df.head(3))
print(df.describe())##描述性分析
print(df.corr())##各特征相关性分析
##计算每行每列数据的缺失值个数
defnum_missing(x):
return sum(x.isnull())
print("Missing values per column:")
print(df.apply(num_missing, axis=0)) #axis=0代表函数应用于每一列
print(" Missing values per row:")
print(df.apply(num_missing, axis=1).head()) #axis=1代表函数应用于每一行"""