爱数课实验 | 第九期-利用机器学习方法进行健康智能诊断

import pandas as pd
import numpy as np

# 不显示警告
import warnings
warnings.filterwarnings('ignore')

data = pd.read_csv("./dataset/Indian Liver Patient Dataset (ILPD).csv")
data.head()

# 查看数据信息
data.info()

data.describe(include='all')

from sklearn.preprocessing import LabelEncoder
# 数字编码
le = LabelEncoder()

# 数字编码
data["Gender"] = le.fit_transform(data["Gender"][:])

data

from sklearn.impute import SimpleImputer

# 采用均值填补法
imp = SimpleImputer(strategy="mean")
data["A/G Ratio"] = imp.fit_transform(data["A/G Ratio"].to_frame())

data.describe()

# 定义label列
calculate_col = "label"
calculate_value = 2

# 将label值从{1，2}，转化为{1，0}并更名为label_cal
data[calculate_col + '_cal'] = calculate_value - data[calculate_col]
# 删除label列
data = data.drop(labels='label', axis=1)

#查看转化后的数据集
data

import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline
%config InlineBackend.figure_format = 'svg'
plt.rcParams['font.sans-serif'] = ['SimHei']  # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False  # 用来正常显示负号

# 定义genders列
genders = ['女性', '男性']
fig, [ax1, ax2] = plt.subplots(1, 2, figsize=(12, 5))

# 画出年龄分布直方图
sns.histplot(data['Age'], kde=False, palette="Set2", ax=ax1)
# 画出性别分布柱状图
sns.countplot(x='Gender', data=data, palette="Set2", ax=ax2)

ax1.set(xlabel='年龄', ylabel='数量')
ax2.set(xlabel='性别', ylabel='数量')
ax2.set_xticklabels(genders)

# 定义两个图像
fig, [ax1, ax2] = plt.subplots(1, 2, figsize=(12, 5))

# 绘制年龄箱线图
sns.boxplot(x='label_cal', y='Age', hue='label_cal', data=data, palette="Set2", ax=ax1)
# 绘制性别柱状图
sns.countplot(x='Gender', hue='label_cal', data=data, palette="Set2", ax=ax2)

# 给图像加标题
ax1.set_title('年龄与患病情况的分布', fontsize=13)
ax2.set_title('性别与患病情况的分布', fontsize=13)

# 给图像添加标签
labels = ['不患病', '患病']
ax1.set(xlabel='是否患病', ylabel='年龄')
ax1.set_xticklabels(labels)
ax1.legend_.remove()
ax2.set(xlabel='性别', ylabel='数量')
ax2.set_xticklabels(genders)
ax2.legend(['不患病', '患病'])

plt.show()

def correlation_heatmap(df):
    hm, ax = plt.subplots(figsize=(12, 8))

# 绘制热力图
    hm = sns.heatmap(
        df.corr(),
        cmap='Blues',
        square=True,
        cbar_kws={'shrink': .9},
        ax=ax,
        annot=True,
        linewidths=0.1, vmax=1.0, linecolor='white',
        annot_kws={'fontsize': 12}
    )

    plt.title('连续型对象间的皮尔森相关系数', y=1.05, size=15)

# 删掉Gender与label_cal列
df = data.drop(['Gender', 'label_cal'], axis=1)
correlation_heatmap(df)

from sklearn.feature_selection import SelectKBest
from sklearn.feature_selection import chi2

# 定义两个新的数据框
gender = pd.DataFrame(data['Gender'])
label_cal = pd.DataFrame(data['label_cal'])

# 使用卡方检验
model1 = SelectKBest(chi2, k=1) # 选择k个最佳特征
model1.fit_transform(gender, label_cal)

# 打印得分
print('性别变量与是否得病之间的得分为：%.4f' % model1.scores_)
print('性别变量与是否得病之间的pvalue为：%.4f' % model1.pvalues_)

from sklearn.feature_selection import f_classif


fdata = pd.DataFrame(data.drop(['Gender', 'label_cal'], axis=1))# 删除gender列与label_cal列
label_cal = pd.DataFrame(data['label_cal'])

# 使用f_classif计算label_cal与其他连续型变量间的关系
F, p_val = f_classif(fdata, label_cal)
# f分布的0.05分位数
print('各连续型变量的名称：')
print(fdata.columns.tolist())
print('各连续型变量与是否得病之间的F值为：')
print(F)
print('各连续型变量与是否得病之间的pvalue为：')
print(p_val)

from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score
x = data.drop(['label_cal'], axis=1) # x为删除label_cal的数据集
y = data['label_cal'] # y为label_cal列

# 划分数据集
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2, random_state=2, stratify=data['label_cal'])

from sklearn.linear_model import LogisticRegression

# 为权重赋值
weights = {0: 1, 1: 1.3}

# 进行logistic回归
lr = LogisticRegression(penalty='l2', random_state=8, class_weight=weights)
lr.fit(x_train, y_train)

# 对y进行预测
y_predprb = lr.predict_proba(x_test)[:, 1]
y_pred = lr.predict(x_test)

from sklearn import metrics
from sklearn.metrics import auc

# 计算fpr，tpr及thresholds的值
fpr, tpr, thresholds = metrics.roc_curve(y_test, y_predprb)
# 计算gmean的值
gmean = np.sqrt(tpr*(1-fpr))

# 计算最大的gmean值对应的thresholds值
dictionary = dict(zip(thresholds,gmean))
max_thresholds = max(dictionary, key=dictionary.get)

print("最大的GMean值为：%.4f"%(max(gmean)))
print("最大的GMean对应的thresholds为：%.4f"%(max_thresholds))

from sklearn.metrics import roc_auc_score
# 计算AUC值
test_roc_auc = roc_auc_score(y_test, y_predprb)
print(test_roc_auc)

# 打印模型分类预测报告
print(classification_report(y_test, y_pred))

# 画出混淆矩阵热力图
cm1 = confusion_matrix(y_test, y_pred)
plt.figure(figsize=(8, 6))
sns.heatmap(cm1, annot=True, linewidths=.5, square=True, cmap='Blues')
plt.ylabel('Actual label')
plt.xlabel('Predicted label')
all_sample_title = 'ROC AUC Score: {0}'.format(round(test_roc_auc,2))
plt.title(all_sample_title, size=15)

from sklearn.tree import DecisionTreeClassifier
from sklearn import tree

# 建立决策树模型
model = DecisionTreeClassifier(random_state=5, class_weight=weights)
model = model.fit(x_train, y_train)

# 对y进行预测
y_predict = model.predict(x_test)
y_predprb = model.predict_proba(x_test)[:, 1]

# 计算AUC值
test_roc_auc = roc_auc_score(y_test, y_predprb)
print(test_roc_auc)

# 打印模型分类预测报告
print(classification_report(y_test, y_predict))

# 绘制混淆矩阵热力图
cm2 = confusion_matrix(y_test, y_predict)
plt.figure(figsize=(9, 9))
sns.heatmap(cm2, annot=True, linewidths=.5, square=True, cmap='Blues')
plt.ylabel('Actual label')
plt.xlabel('Predicted label')
all_sample_title = 'ROC AUC score: {0}'.format(round(test_roc_auc,2))
plt.title(all_sample_title, size=15)

from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import cross_val_score

# 建立随机森林模型
ran_for = RandomForestClassifier(n_estimators=80, random_state=0, class_weight=weights)
ran_for.fit(x_train, y_train)

# 对y进行预测
y_pred_ran = ran_for.predict(x_test)
y_predprb = ran_for.predict_proba(x_test)[:, 1]

# 计算AUC值
test_roc_auc = roc_auc_score(y_test, y_predprb)
print(test_roc_auc)

# 打印模型分类预测报告
print(classification_report(y_test, y_pred_ran, digits=2))

# 绘制混淆矩阵热力图
cm3 = confusion_matrix(y_test, y_pred_ran)
plt.figure(figsize=(9, 9))
sns.heatmap(cm3, annot=True, linewidths=.5, square=True, cmap='Blues')
plt.ylabel('Actual label')
plt.xlabel('Predicted label')
all_sample_title = 'ROC AUC score: {0}'.format(round(test_roc_auc,2))
plt.title(all_sample_title, size=15)

from sklearn import preprocessing
X = data.iloc[:, :-1] # 除了label_cal列
y = data['label_cal']

np.random.seed(123)
perm = np.random.permutation(len(X)) # 将数组随机生成一个新的序列
X = X.loc[perm]
y = y[perm]
X = preprocessing.scale(X)# 进行标准化处理

from sklearn.decomposition import PCA
# 使用PCA进行降维
pca = PCA(copy=True, n_components=6, whiten=False, random_state=1)
X_new = pca.fit_transform(X)

print(u'所保留的6个主成分的方差贡献率为：')
print(pca.explained_variance_ratio_)
print(u'排名前2的主成分特征向量为：')
print(pca.components_[0:1])
print(u'累计方差贡献率为：')
print(sum(pca.explained_variance_ratio_))

# 对数据集进行划分
x_train, x_test, y_train, y_test = train_test_split(X_new, y, test_size=0.2, random_state=2, stratify=data['label_cal'])

# 建立随机森林模型
ran_for = RandomForestClassifier(n_estimators=80, random_state=0, class_weight=weights)
# 训练模型
ran_for.fit(x_train, y_train)

# 对y进行预测
y_pred_ran = ran_for.predict(x_test)
y_predprb = ran_for.predict_proba(x_test)[:, 1]

# 计算AUC值
test_roc_auc = roc_auc_score(y_test, y_predprb)
print(test_roc_auc)

# 打印模型分类预测报告
print(classification_report(y_test, y_pred_ran, digits=2))

# 绘制混淆矩阵热力图
cm4 = confusion_matrix(y_test, y_pred_ran)
plt.figure(figsize=(9, 9))
sns.heatmap(cm4, annot=True, linewidths=.5, square=True, cmap='Blues')
plt.ylabel('Actual label')
plt.xlabel('Predicted label')
all_sample_title = 'ROC AUC score: {0}'.format(round(test_roc_auc,2))
plt.title(all_sample_title, size=15)