机器学习奥林匹克-身体健康与幸福之心脏病预测

import numpy as np
import pandas as pd 
import matplotlib.pyplot as plt
%matplotlib inline
import seaborn as sns
sns.set_style('whitegrid')
import warnings
warnings.filterwarnings("ignore")

# 解压缩数据
# !unzip data/data127426/ml-olympiad-good-health-and-well-being.zip

# 读取数据
sub = pd.read_csv('sample_submission.csv')
test = pd.read_csv('test.csv')
train = pd.read_csv('train.csv')

train.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 177576 entries, 0 to 177575
Data columns (total 23 columns):
 #   Column             Non-Null Count   Dtype
---  ------             --------------   -----
 0   PatientID          177576 non-null  int64
 1   HighBP             177576 non-null  int64
 2   HighChol           177576 non-null  int64
 3   CholCheck          177576 non-null  int64
 4   BMI                177576 non-null  int64
 5   Smoker             177576 non-null  int64
 6   Stroke             177576 non-null  int64
 7   Diabetes           177576 non-null  int64
 8   PhysActivity       177576 non-null  int64
 9   Fruits             177576 non-null  int64
 10  Veggies            177576 non-null  int64
 11  HvyAlcoholConsump  177576 non-null  int64
 12  AnyHealthcare      177576 non-null  int64
 13  NoDocbcCost        177576 non-null  int64
 14  GenHlth            177576 non-null  int64
 15  MentHlth           177576 non-null  int64
 16  PhysHlth           177576 non-null  int64
 17  DiffWalk           177576 non-null  int64
 18  Sex                177576 non-null  int64
 19  Age                177576 non-null  int64
 20  Education          177576 non-null  int64
 21  Income             177576 non-null  int64
 22  target             177576 non-null  int64
dtypes: int64(23)
memory usage: 31.2 MB

train.head()
.dataframe tbody tr th:only-of-type {         vertical-align: middle;     } .dataframe tbody tr th {     vertical-align: top; } .dataframe thead th {     text-align: right; }

train.describe()
.dataframe tbody tr th:only-of-type {         vertical-align: middle;     } .dataframe tbody tr th {     vertical-align: top; } .dataframe thead th {     text-align: right; }

print(train.shape)
print(test.shape)

(177576, 23)
(76104, 22)

train.isnull().sum()

PatientID            0
HighBP               0
HighChol             0
CholCheck            0
BMI                  0
Smoker               0
Stroke               0
Diabetes             0
PhysActivity         0
Fruits               0
Veggies              0
HvyAlcoholConsump    0
AnyHealthcare        0
NoDocbcCost          0
GenHlth              0
MentHlth             0
PhysHlth             0
DiffWalk             0
Sex                  0
Age                  0
Education            0
Income               0
target               0
dtype: int64

test.isnull().sum()

PatientID            0
HighBP               0
HighChol             0
CholCheck            0
BMI                  0
Smoker               0
Stroke               0
Diabetes             0
PhysActivity         0
Fruits               0
Veggies              0
HvyAlcoholConsump    0
AnyHealthcare        0
NoDocbcCost          0
GenHlth              0
MentHlth             0
PhysHlth             0
DiffWalk             0
Sex                  0
Age                  0
Education            0
Income               0
dtype: int64

train.duplicated().any()

False

test.duplicated().any()

False

train.hist(figsize=(20,12));

train[['BMI','Age', 'Income']].hist(figsize=(20,12));

# 各类值数量
train.nunique()

PatientID            177576
HighBP                    2
HighChol                  2
CholCheck                 2
BMI                      81
Smoker                    2
Stroke                    2
Diabetes                  3
PhysActivity              2
Fruits                    2
Veggies                   2
HvyAlcoholConsump         2
AnyHealthcare             2
NoDocbcCost               2
GenHlth                   5
MentHlth                 31
PhysHlth                 31
DiffWalk                  2
Sex                       2
Age                      13
Education                 6
Income                    8
target                    2
dtype: int64

plt.figure(figsize=(20,8))
plt.subplot(2,4,1)
sns.countplot(train.HighBP)
plt.subplot(2,4,2)
sns.countplot(train.HighChol)
plt.subplot(2,4,3)
sns.countplot(train.CholCheck)
plt.subplot(2,4,4)
sns.countplot(train.Smoker)
plt.subplot(2,4,5)
sns.countplot(train.Stroke)
plt.subplot(2,4,6)
sns.countplot(train.Sex)
plt.subplot(2,4,7)
sns.countplot(train.Education)
plt.subplot(2,4,8)
sns.countplot(train.target);

# 协相关
train.corr()
.dataframe tbody tr th:only-of-type {         vertical-align: middle;     } .dataframe tbody tr th {     vertical-align: top; } .dataframe thead th {     text-align: right; }

# 热力图
plt.figure(figsize=(20,12))
sns.heatmap(train.corr(), annot=True)

<matplotlib.axes._subplots.AxesSubplot at 0x7f845ce2eed0>

train.columns

Index(['PatientID', 'HighBP', 'HighChol', 'CholCheck', 'BMI', 'Smoker',       'Stroke', 'Diabetes', 'PhysActivity', 'Fruits', 'Veggies',       'HvyAlcoholConsump', 'AnyHealthcare', 'NoDocbcCost', 'GenHlth',       'MentHlth', 'PhysHlth', 'DiffWalk', 'Sex', 'Age', 'Education', 'Income',       'target'],
      dtype='object')

x = train.drop(['PatientID', 'target'], axis=1)
y=train['target']

from sklearn.model_selection import train_test_split
x_train, x_test, y_train,  y_test = train_test_split(x,y, test_size=.3, random_state=42)

from sklearn.linear_model import LogisticRegression

lr = LogisticRegression(solver='liblinear')

lr.fit(x_train,y_train)

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='liblinear', tol=0.0001, verbose=0,
                   warm_start=False)

# Base
from sklearn.metrics import classification_report, confusion_matrix, f1_score
print('Training score:', lr.score(x_train, y_train))
print('*' *20)
print('Test score:',lr.score(x_test, y_test))
print('*' *20)
print('f1_score:', f1_score(y_test, lr.predict(x_test)))
print('*' *20)
print(confusion_matrix(y_test, lr.predict(x_test)))
print('*' *20)
print(classification_report(y_test, lr.predict(x_test)))

Training score: 0.9081679444582995
********************
Test score: 0.9067820471908847
********************
f1_score: 0.19695989650711515
********************
[[47698   478]
 [ 4488   609]]
********************
              precision    recall  f1-score   support
           0       0.91      0.99      0.95     48176
           1       0.56      0.12      0.20      5097
    accuracy                           0.91     53273
   macro avg       0.74      0.55      0.57     53273
weighted avg       0.88      0.91      0.88     53273

testt = test.drop('PatientID', axis=1)
pred = lr.predict(testt)
subm = pd.DataFrame({'PatientID':sub.PatientID, 'target':pred}, index=None) # convert to dataframe
subm.to_csv('first_submission.csv', index=False) # write it to a csv file