import numpy as np
import matplotlib.pyplot as plt
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import PolynomialFeatures
from sklearn.linear_model import LinearRegression
from sklearn.model_selection import cross_val_score
%matplotlib inline
def true_fun (X) :return np.cos(1.5 * np.pi * X)
np.random.seed(0 )
n_samples = 30
degrees = [1 , 4 , 15 ]
X = np.sort(np.random.rand(n_samples))
y = true_fun(X) + np.random.randn(n_samples) * 0.1
print(X.shape,y.shape)
print()
print(X)
print()
print(y)
(30,) (30,)
[0.0202184 0.07103606 0.0871293 0.11827443 0.14335329 0.38344152
0.41466194 0.4236548 0.43758721 0.46147936 0.52184832 0.52889492
0.54488318 0.5488135 0.56804456 0.60276338 0.63992102 0.64589411
0.71518937 0.77815675 0.78052918 0.79172504 0.79915856 0.83261985
0.87001215 0.891773 0.92559664 0.94466892 0.96366276 0.97861834]
[ 1.0819082 0.87027612 1.14386208 0.70322051 0.78494746 -0.25265944
-0.22066063 -0.26595867 -0.4562644 -0.53001927 -0.86481449 -0.99462675
-0.87458603 -0.83407054 -0.77090649 -0.83476183 -1.03080067 -1.02544303
-1.0788268 -1.00713288 -1.03009698 -0.63623922 -0.86230652 -0.75328767
-0.70023795 -0.41043495 -0.50486767 -0.27907117 -0.25994628 -0.06189804]
degree=5
polynomial_features = PolynomialFeatures(degree=degree,include_bias=False )
linear_regression = LinearRegression()
pipeline = Pipeline([("polynomial_features" , polynomial_features),("linear_regression" , linear_regression)])
pipeline.fit(X[:, np.newaxis], y)
scores = cross_val_score(pipeline, X[:, np.newaxis], y,scoring="neg_mean_squared_error" , cv=10 )
X_test = np.linspace(0 , 1 , 100 )
plt.plot(X_test, pipeline.predict(X_test[:, np.newaxis]), label="Model" )
plt.plot(X_test, true_fun(X_test), label="True function" )
plt.scatter(X, y, edgecolor='b' , s=20 , label="Samples" )
plt.xlabel("x" )
plt.ylabel("y" )
plt.xlim((0 , 1 ))
plt.ylim((-2 , 2 ))
plt.legend(loc="best" )
plt.title("Degree {}\nMSE = {:.2e}(+/- {:.2e})" .format(degree, -scores.mean(), scores.std()))
Text(0.5, 1.0, 'Degree 5\nMSE = 9.87e-02(+/- 2.29e-01)')
plt.figure(figsize=(14 , 5 ))
for i in range(len(degrees)):
ax = plt.subplot(1 , len(degrees), i + 1 )
plt.setp(ax, xticks=(), yticks=())
polynomial_features = PolynomialFeatures(degree=degrees[i],include_bias=False )
linear_regression = LinearRegression()
pipeline = Pipeline([("polynomial_features" , polynomial_features),
("linear_regression" , linear_regression)])
pipeline.fit(X[:, np.newaxis], y)
scores = cross_val_score(pipeline, X[:, np.newaxis], y,
scoring="neg_mean_squared_error" , cv=10 )
X_test = np.linspace(0 , 1 , 100 )
plt.plot(X_test, pipeline.predict(X_test[:, np.newaxis]), label="Model" )
plt.plot(X_test, true_fun(X_test), label="True function" )
plt.scatter(X, y, edgecolor='b' , s=20 , label="Samples" )
plt.xlabel("x" )
plt.ylabel("y" )
plt.xlim((0 , 1 ))
plt.ylim((-2 , 2 ))
plt.legend(loc="best" )
plt.title("Degree {}\nMSE = {:.2e}(+/- {:.2e})" .format(degrees[i], -scores.mean(), scores.std()))
plt.show()
分类器
导入必要的包、读入数据
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from pandas import Series, DataFrame
from scipy import stats
%matplotlib inline
mydata=pd.read_excel('./CLFData/个人收入水平调查分析.xlsx' )
mydata.head(100 )
年龄
受教育时间
性别
资产净增
资产损失
一周工作时间
收入水平
0
39
13
Male
2174
0
40
<=50K
1
50
13
Male
0
0
13
<=50K
2
38
9
Male
0
0
40
<=50K
3
53
7
Male
0
0
40
<=50K
4
28
13
Female
0
0
40
<=50K
...
...
...
...
...
...
...
...
95
29
10
Male
0
0
50
<=50K
96
48
16
Male
0
1902
60
>50K
97
37
10
Male
0
0
48
>50K
98
48
12
Female
0
0
40
<=50K
99
32
9
Male
0
0
40
<=50K
100 rows × 7 columns
检查数据缺失度
mydata.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 32561 entries, 0 to 32560
Data columns (total 7 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 年龄 32561 non-null int64
1 受教育时间 32561 non-null int64
2 性别 32561 non-null object
3 资产净增 32561 non-null int64
4 资产损失 32561 non-null int64
5 一周工作时间 32561 non-null int64
6 收入水平 32561 non-null object
dtypes: int64(5), object(2)
memory usage: 1.7+ MB
变量类型处理
mydata.describe()
mydata['性别' ]=mydata['性别' ].map( {'Female' : 1 , 'Male' : 0 } ).astype(int)
mydata['收入水平' ]=mydata['收入水平' ].map( {'>50K' : 1 , '<=50K' : 0 } ).astype(int)
mydata.head()
年龄
受教育时间
性别
资产净增
资产损失
一周工作时间
收入水平
0
39
13
0
2174
0
40
0
1
50
13
0
0
0
13
0
2
38
9
0
0
0
40
0
3
53
7
0
0
0
40
0
4
28
13
1
0
0
40
0
分割自变量和因变量
X=mydata.drop(['收入水平' ],axis=1 )
X
年龄
受教育时间
性别
资产净增
资产损失
一周工作时间
0
39
13
0
2174
0
40
1
50
13
0
0
0
13
2
38
9
0
0
0
40
3
53
7
0
0
0
40
4
28
13
1
0
0
40
...
...
...
...
...
...
...
32556
27
12
1
0
0
38
32557
40
9
0
0
0
40
32558
58
9
1
0
0
40
32559
22
9
0
0
0
20
32560
52
9
1
15024
0
40
32561 rows × 6 columns
y=mydata['收入水平' ]
X.shape,y.shape
((32561, 6), (32561,))
from sklearn.model_selection import train_test_split
X_train,X_test,y_train,y_test=train_test_split(X,y,random_state=4 )
X_train.shape,X_test.shape,y_train.shape,y_test.shape
((24420, 6), (8141, 6), (24420,), (8141,))
决策树C4.5
from sklearn import tree
clf=tree.DecisionTreeClassifier()
clf.fit(X,y)
print(clf.predict(X))
print('准确率：' ,1 -(clf.predict(X)!=y).sum()/len(y))
y_pred_dt=clf.predict(X)
from sklearn import metrics
print(metrics.classification_report(y,y_pred_dt))
print(metrics.confusion_matrix(y,y_pred_dt))
print(metrics.f1_score(y,y_pred_dt))
[0 0 0 ... 0 0 1]
准确率： 0.8955806025613464
precision recall f1-score support
0 0.90 0.97 0.93 24720
1 0.89 0.65 0.75 7841
accuracy 0.90 32561
macro avg 0.89 0.81 0.84 32561
weighted avg 0.90 0.90 0.89 32561
[[24097 623]
[ 2777 5064]]
0.7486694263749261
from sklearn import tree
clf=tree.DecisionTreeClassifier()
clf.fit(X_train,y_train)
print(clf.predict(X_train))
print('准确率：' ,1 -(clf.predict(X_train)!=y_train).sum()/len(y_train))
y_test_pred_dt=clf.predict(X_test)
from sklearn import metrics
print(metrics.classification_report(y_test,y_test_pred_dt))
print(metrics.confusion_matrix(y_test,y_test_pred_dt))
print(metrics.f1_score(y_test,y_test_pred_dt))
[0 0 0 ... 0 1 0]
准确率： 0.8997952497952498
precision recall f1-score support
0 0.86 0.91 0.88 6172
1 0.66 0.51 0.58 1969
accuracy 0.82 8141
macro avg 0.76 0.71 0.73 8141
weighted avg 0.81 0.82 0.81 8141
[[5647 525]
[ 957 1012]]
0.5772960638904735
KNN
from sklearn.neighbors import KNeighborsClassifier
clf=KNeighborsClassifier(n_neighbors=3 )
clf.fit(X,y)
print(clf.predict(X))
print('准确率：' ,1 -(clf.predict(X)!=y).sum()/len(y))
y_pred_knn=clf.predict(X)
from sklearn import metrics
print(metrics.classification_report(y,y_pred_knn))
print(metrics.confusion_matrix(y,y_pred_knn))
print(metrics.f1_score(y,y_pred_knn))
[0 0 1 ... 0 0 1]
准确率： 0.8591259482202636
precision recall f1-score support
0 0.89 0.93 0.91 24720
1 0.74 0.64 0.69 7841
accuracy 0.86 32561
macro avg 0.82 0.78 0.80 32561
weighted avg 0.85 0.86 0.86 32561
[[22947 1773]
[ 2814 5027]]
0.6867017280240421
from sklearn.neighbors import KNeighborsClassifier
clf=KNeighborsClassifier(n_neighbors=3 )
clf.fit(X_train,y_train)
print(clf.predict(X_train))
print('准确率：' ,1 -(clf.predict(X_train)!=y_train).sum()/len(y_train))
y_test_pred_knn=clf.predict(X_test)
from sklearn import metrics
print(metrics.classification_report(y_test,y_test_pred_knn))
print(metrics.confusion_matrix(y_test,y_test_pred_knn))
print(metrics.f1_score(y_test,y_test_pred_knn))
[0 0 0 ... 0 1 0]
准确率： 0.8949356592242252
precision recall f1-score support
0 0.87 0.90 0.88 6172
1 0.63 0.56 0.60 1969
accuracy 0.82 8141
macro avg 0.75 0.73 0.74 8141
weighted avg 0.81 0.82 0.81 8141
[[5533 639]
[ 858 1111]]
0.5974724388276419
逻辑回归
from sklearn.linear_model import LogisticRegression
clf=LogisticRegression(max_iter=200 )
clf.fit(X,y)
print(clf.predict(X))
print('准确率：' ,1 -(clf.predict(X)!=y).sum()/len(y))
y_pred_log=clf.predict(X)
from sklearn import metrics
print(metrics.classification_report(y,y_pred_log))
print(metrics.confusion_matrix(y,y_pred_log))
print(metrics.f1_score(y,y_pred_log))
[1 0 0 ... 0 0 1]
准确率： 0.8225177359417708
precision recall f1-score support
0 0.84 0.95 0.89 24720
1 0.72 0.43 0.54 7841
accuracy 0.82 32561
macro avg 0.78 0.69 0.72 32561
weighted avg 0.81 0.82 0.81 32561
[[23384 1336]
[ 4443 3398]]
0.5404373757455269
from sklearn.linear_model import LogisticRegression
clf=LogisticRegression(max_iter=200 )
clf.fit(X_train,y_train)
print(clf.predict(X_train))
print('准确率：' ,1 -(clf.predict(X_train)!=y_train).sum()/len(y_train))
y_test_pred_log=clf.predict(X_test)
from sklearn import metrics
print(metrics.classification_report(y_test,y_test_pred_log))
print(metrics.confusion_matrix(y_test,y_test_pred_log))
print(metrics.f1_score(y_test,y_test_pred_log))
[0 0 0 ... 0 1 0]
准确率： 0.8215397215397215
precision recall f1-score support
0 0.84 0.94 0.89 6172
1 0.72 0.45 0.56 1969
accuracy 0.82 8141
macro avg 0.78 0.70 0.72 8141
weighted avg 0.81 0.82 0.81 8141
[[5822 350]
[1075 894]]
0.5564892623716153
clf
LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,
intercept_scaling=1, l1_ratio=None, max_iter=100,
multi_class='auto', n_jobs=None, penalty='l2',
random_state=None, solver='lbfgs', tol=0.0001, verbose=0,
warm_start=False)
"""
y_pred=clf.predict(X)
from sklearn import metrics
print(metrics.classification_report(y,y_pred))
print(metrics.confusion_matrix(y,y_pred))
"""
from sklearn.metrics import roc_curve, auc,confusion_matrix,precision_score ,recall_score,f1_score
predictions = clf.predict_proba(X_test)
false_positive_rate, recall, thresholds = roc_curve(y, predictions[:, 1 ])
roc_auc = auc(false_positive_rate, recall)
plt.title('Receiver Operating Characteristic' )
plt.plot(false_positive_rate, recall, 'b' , label='AUC = %0.2f' % roc_auc)
plt.legend(loc='lower right' )
plt.plot([0 , 1 ], [0 , 1 ], 'r--' )
plt.xlim([0.0 , 1.0 ])
plt.ylim([0.0 , 1.0 ])
plt.ylabel('Recall' )
plt.xlabel('Fall-out' )
plt.show()
