[人工智能]【项目实战】心脏病患者数据分析和建模

导入科学计算库

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import warnings
warnings.filterwarnings('ignore')
from pylab import mpl
# 设置显示中文字体
mpl.rcParams["font.sans-serif"] = ['SimHei']
# 设置正常显示符号
mpl.rcParams["axes.unicode_minus"] = False

读取数据

heart_df = pd.read_csv("./data/heart.csv")
heart_df.head()

heart_df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 303 entries, 0 to 302
Data columns (total 14 columns):
 #   Column    Non-Null Count  Dtype  
---  ------    --------------  -----  
 0   age       303 non-null    int64  
 1   sex       303 non-null    int64  
 2   cp        303 non-null    int64  
 3   trestbps  303 non-null    int64  
 4   chol      303 non-null    int64  
 5   fbs       303 non-null    int64  
 6   restecg   303 non-null    int64  
 7   thalach   303 non-null    int64  
 8   exang     303 non-null    int64  
 9   oldpeak   303 non-null    float64
 10  slope     303 non-null    int64  
 11  ca        303 non-null    int64  
 12  thal      303 non-null    int64  
 13  target    303 non-null    int64  
dtypes: float64(1), int64(13)
memory usage: 33.3 KB

• age - 年龄
• sex - (1 = male(男性); 0 = (女性))
• cp - chest pain type(胸部疼痛类型)（1：典型的心绞痛-typical，2：非典型心绞痛-atypical，3：没有心绞痛-non-anginal，4：无症状-asymptomatic）
• trestbps - 静息血压 (in mm Hg on admission to the hospital)
• chol - 胆固醇 in mg/dl
• fbs - (空腹血糖 > 120 mg/dl) (1 = true; 0 = false)
• restecg - 静息心电图测量（0：普通，1：ST-T波异常，2：可能左心室肥大）
• thalach - 最高心跳率
• exang - 运动诱发心绞痛 (1 = yes; 0 = no)
• oldpeak - 运动相对于休息引起的ST抑制
• slope - 运动ST段的峰值斜率（1：上坡-upsloping，2：平的-flat，3：下坡-downsloping）
• ca - 主要血管数目(0-4)
• thal - 一种叫做地中海贫血的血液疾病（3 = normal; 6 = 固定的缺陷-fixed defect; 7 = 可逆的缺陷-reversable defect）
• target - 是否患病 (1=yes, 0=no)

患病的分布情况

fig,axes = plt.subplots(1,2,figsize=(10,5), dpi=100)
ax = heart_df.target.value_counts().plot(kind="bar",ax=axes[0])
ax.set_title("患病分布")
ax.set_xlabel("1：患病，0：未患病")

heart_df.target.value_counts().plot(kind="pie",autopct="%.2f%%",labels=['患病','未患病'],ax=axes[1])
plt.show()

性别和患病的分布

plt.figure(figsize=(10,5), dpi=100)
ax1 = plt.subplot(121)
ax = sns.countplot(x="sex",hue='target',data=heart_df,ax=ax1)
ax.set_xlabel("0：女性，1：男性")

ax2 = plt.subplot(222)
heart_df[heart_df['target'] == 0].sex.value_counts().plot(kind="pie",autopct="%.2f%%",labels=['男性','女性'],ax=ax2)
ax2.set_title("未患病性别比例")

ax2 = plt.subplot(224)
heart_df[heart_df['target'] == 1].sex.value_counts().plot(kind="pie",autopct="%.2f%%",labels=['男性','女性'],ax=ax2)
ax2.set_title("患病性别比例")

Text(0.5, 1.0, '患病性别比例')

年龄分布情况

fig,axes = plt.subplots(2,1,figsize=(20,10), dpi=100)
ax = sns.countplot(x="age",hue="target",data=heart_df,ax=axes[0])
ax.set_xlabel("0：女性，1：男性")

# 0-45：青年人，45-59：中年人，60-100：老年人
age_type = pd.cut(heart_df.age,bins=[0,45,60,100],include_lowest=True,right=False,labels=['青年人','中年人','老年人'])
age_target_df = pd.concat([age_type,heart_df.target],axis=1)
sns.countplot(x="age",hue='target',data=age_target_df)
plt.show()

统一看下所有特征的分布情况

fig,axes = plt.subplots(7,2,figsize=(10,20), dpi=100)
for x in range(0,14):
    plt.subplot(7,2,x+1)
    sns.distplot(heart_df.iloc[:,x],kde=True)

plt.tight_layout()

# 绘制热力图
plt.figure(figsize=(10,6), dpi=100)
sns.heatmap(heart_df.corr(),cmap="Blues",annot=True)
plt.show()

建模操作

数据预处理

features = heart_df.drop(columns=['target'])
targets = heart_df['target']

features.head()

# 将离散型数据，从普通的0,1,2这些，转换成真正的字符串表示

# sex
features.loc[features['sex']==0,'sex'] = 'female'
features.loc[features['sex']==1,'sex'] = 'male'

# cp
features.loc[features['cp'] == 1,'cp'] = 'typical'
features.loc[features['cp'] == 2,'cp'] = 'atypical'
features.loc[features['cp'] == 3,'cp'] = 'non-anginal'
features.loc[features['cp'] == 4,'cp'] = 'asymptomatic'

# fbs
features.loc[features['fbs'] == 1,'fbs'] = 'true'
features.loc[features['fbs'] == 0,'fbs'] = 'false'

# exang
features.loc[features['exang'] == 1,'exang'] = 'true'
features.loc[features['exang'] == 0,'exang'] = 'false'

# slope
features.loc[features['slope'] == 1,'slope'] = 'true'
features.loc[features['slope'] == 2,'slope'] = 'true'
features.loc[features['slope'] == 3,'slope'] = 'true'

# thal
features.loc[features['thal'] == 3,'thal'] = 'normal'
features.loc[features['thal'] == 3,'thal'] = 'fixed'
features.loc[features['thal'] == 3,'thal'] = 'reversable'

# restecg
# 0：普通，1：ST-T波异常，2：可能左心室肥大
features.loc[features['restecg'] == 0,'restecg'] = 'normal'
features.loc[features['restecg'] == 1,'restecg'] = 'ST-T abnormal'
features.loc[features['restecg'] == 2,'restecg'] = 'Left ventricular hypertrophy'

# ca
features['ca'].astype("object")

# thal
features.thal.astype("object")

features.head()

from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split

features = pd.get_dummies(features)  # 文本特征onehot表示
features_temp = StandardScaler().fit_transform(features)
# features_temp = StandardScaler().fit_transform(pd.get_dummies(features))

X_train,X_test,y_train,y_test = train_test_split(features_temp,targets,test_size=0.25)

1. K近邻
2. 决策树
3. 随机森林
4. 逻辑回归
5. SGD分类

K近邻

from sklearn.neighbors import KNeighborsClassifier
from sklearn.model_selection import cross_val_score
from sklearn.metrics import precision_score,recall_score,f1_score
from sklearn.metrics import precision_recall_curve,roc_curve,average_precision_score,auc
# https://www.jianshu.com/p/c61ae11cc5f6

def plotting(estimator,y_test):
    fig,axes = plt.subplots(1,2,figsize=(10,5), dpi=100)
    y_predict_proba = estimator.predict_proba(X_test)
    precisions,recalls,thretholds = precision_recall_curve(y_test,y_predict_proba[:,1])
    axes[0].plot(precisions,recalls)
    axes[0].set_title("平均精准率：%.2f"%average_precision_score(y_test,y_predict_proba[:,1]))
    axes[0].set_xlabel("召回率")
    axes[0].set_ylabel("精准率")

    fpr,tpr,thretholds = roc_curve(y_test,y_predict_proba[:,1])
    axes[1].plot(fpr,tpr)
    axes[1].set_title("AUC值：%.2f"%auc(fpr,tpr))
    axes[1].set_xlabel("FPR")
    axes[1].set_ylabel("TPR")

# 1. K近邻
knn = KNeighborsClassifier(n_neighbors=5)
scores = cross_val_score(knn,features_temp,targets,cv=5)
print("准确率：",scores.mean())

knn.fit(X_train,y_train)

y_predict = knn.predict(X_test)
# 精准率
print("精准率：",precision_score(y_test,y_predict))
# 召回率
print("召回率：",recall_score(y_test,y_predict))
# F1-Score
print("F1得分：",f1_score(y_test,y_predict))

plotting(knn,y_test)

准确率： 0.8317486338797814
精准率： 0.8260869565217391
召回率： 0.8837209302325582
F1得分： 0.853932584269663

决策树

from sklearn.tree import DecisionTreeClassifier
tree = DecisionTreeClassifier(max_depth=10)
tree.fit(X_train,y_train)

plotting(tree,y_test)

随机森林

from sklearn.ensemble import RandomForestClassifier
rf = RandomForestClassifier(n_estimators=100)
rf.fit(X_train,y_train)
plotting(rf,y_test)

逻辑回归

from sklearn.linear_model import LogisticRegression
logic = LogisticRegression(tol=1e-10)
logic.fit(X_train,y_train)
plotting(logic,y_test)

SGD分类

from sklearn.linear_model import SGDClassifier
sgd = SGDClassifier(loss="log")
sgd.fit(X_train,y_train)
plotting(sgd,y_test)

plt.figure(figsize=(15, 8), dpi=100)
importances = pd.Series(data=rf.feature_importances_,index=features.columns).sort_values(ascending=False)
sns.barplot(y=importances.index,x=importances.values,orient='h')
plt.show()

加油!

感谢!

努力!