当前位置： 首页 > news >正文

第13周：LSTM实现糖尿病探索与预测

news 来源：原创 2025/5/3 4:55:28
  • 🍨 本文为🔗365天深度学习训练营 中的学习记录博客
  • *🍖 参考：K同学啊

1.数据预处理

1.1硬件环境配置

import torch.nn as nn
import torch.nn.functional as F
import torchvision,torch

# 设置硬件设备，如果有GPU则使用，没有则使用cpu
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
device

device(type='cpu')

1.2数据导入

import numpy             as np
import pandas            as pd
import seaborn           as sns
from sklearn.model_selection   import train_test_split
import matplotlib.pyplot as plt
plt.rcParams['savefig.dpi'] = 500 #图片像素
plt.rcParams['figure.dpi']  = 500 #分辨率

plt.rcParams['font.sans-serif'] = ['SimHei']  # 用来正常显示中文标签

import warnings 
warnings.filterwarnings("ignore")

DataFrame=pd.read_excel('./data/dia.xls')
DataFrame.head()
卡号性别年龄高密度脂蛋白胆固醇低密度脂蛋白胆固醇极低密度脂蛋白胆固醇甘油三酯总胆固醇脉搏舒张压高血压史尿素氮尿酸肌酐体重检查结果是否糖尿病
0180544210381.252.991.070.645.31838304.99243.35010
1180544220311.151.990.840.503.98856304.72391.04710
2180544230271.292.210.690.604.19736105.87325.75110
3180544240330.932.010.660.843.60836002.40203.24020
4180544250361.172.830.830.734.83856704.09236.84300
DataFrame.shape

(1006, 16)

1.3数据检查

# 查看数据是否有缺失值
print('数据缺失值---------------------------------')
print(DataFrame.isnull().sum())

数据缺失值---------------------------------
卡号            0
性别            0
年龄            0
高密度脂蛋白胆固醇     0
低密度脂蛋白胆固醇     0
极低密度脂蛋白胆固醇    0
甘油三酯          0
总胆固醇          0
脉搏            0
舒张压           0
高血压史          0
尿素氮           0
尿酸            0
肌酐            0
体重检查结果        0
是否糖尿病         0
dtype: int64

# 查看数据是否有重复值
print('数据重复值---------------------------------')
print('数据集的重复值为:'f'{DataFrame.duplicated().sum()}')

数据重复值---------------------------------
数据集的重复值为:0

1.4数据分析

# feature_map = {
#     '年龄': '年龄',
#     '高密度脂蛋白胆固醇': '高密度脂蛋白胆固醇',
#     '低密度脂蛋白胆固醇': '低密度脂蛋白胆固醇',
#     '极低密度脂蛋白胆固醇': '极低密度脂蛋白胆固醇',
#     '甘油三酯': '甘油三酯',
#     '总胆固醇': '总胆固醇',
#     '脉搏': '脉搏',
#     '舒张压':'舒张压',
#     '高血压史':'高血压史',
#     '尿素氮':'尿素氮',
#     '尿酸':'尿酸',
#     '肌酐':'肌酐',
#     '体重检查结果':'体重检查结果'
# }
# plt.figure(figsize=(15, 10))

# for i, (col, col_name) in enumerate(feature_map.items(), 1):
#     plt.subplot(3, 5, i)
#     sns.boxplot(x=DataFrame['是否糖尿病'], y=DataFrame[col])
#     plt.title(f'{col_name}的箱线图', fontsize=14)
#     plt.ylabel('数值', fontsize=12)
#     plt.grid(axis='y', linestyle='--', alpha=0.7)

# plt.tight_layout()
# plt.show()
import matplotlib.pyplot as plt
import seaborn as sns

# 创建一个新的图形，并设置其大小为15x10英寸
plt.figure(figsize=(15, 10))

# 遍历feature_map字典中的每个键值对
for i, (col, col_name) in enumerate(feature_map.items(), 1):
    # 创建一个新的子图，该子图位于3行5列的网格的第i个位置
    plt.subplot(3, 5, i)
    
    # 创建一个箱线图，该图显示DataFrame中是否糖尿病列与当前特征列之间的关系
    sns.boxplot(x=DataFrame['是否糖尿病'], y=DataFrame[col], palette="Set1")
    
    # 设置当前子图的标题为当前特征的中文名称加上"的箱线图"，字体大小为14
    plt.title(f'{col_name}的箱线图', fontsize=14)
    
    # 设置当前子图的y轴标签为"数值"，字体大小为12
    plt.ylabel('数值', fontsize=12)
    
    # 在当前子图的y轴上添加网格线，线型为"--"，透明度为0.7
    plt.grid(axis='y', linestyle='--', alpha=0.7)

# 调整子图之间的间距，使得它们不会相互重叠
plt.tight_layout()

# 显示图形
plt.show()

​
在这里插入图片描述

​

1.5相关性分析

import plotly
import plotly.express as px

# 删除列 '卡号'
DataFrame.drop(columns=['卡号'], inplace=True)
# 计算各列之间的相关系数
df_corr = DataFrame.corr()

# 相关矩阵生成函数
def corr_generate(df):
    fig = px.imshow(df,text_auto=True,aspect="auto",color_continuous_scale='RdBu_r')
    fig.show()

# 生成相关矩阵
corr_generate(df_corr)

在这里插入图片描述

2.LSTM模型

2.1划分数据集

from sklearn.preprocessing import StandardScaler

# '高密度脂蛋白胆固醇'字段与糖尿病负相关，故而在 X 中去掉该字段
X = DataFrame.drop(['是否糖尿病','高密度脂蛋白胆固醇'],axis=1)
y = DataFrame['是否糖尿病']

# 数据集标准化处理
sc_X    = StandardScaler()
X = sc_X.fit_transform(X)

X = torch.tensor(np.array(X), dtype=torch.float32)
y = torch.tensor(np.array(y), dtype=torch.int64)

train_X, test_X, train_y, test_y = train_test_split(X, y, 
                                                    test_size=0.2,
                                                    random_state=1)
train_X.shape, train_y.shape

(torch.Size([804, 13]), torch.Size([804]))

2.2数据集构建

from torch.utils.data import TensorDataset, DataLoader

train_dl = DataLoader(TensorDataset(train_X, train_y),
                      batch_size=64, 
                      shuffle=False)

test_dl  = DataLoader(TensorDataset(test_X, test_y),
                      batch_size=64, 
                      shuffle=False)

2.3定义模型

class model_lstm(nn.Module):
    def __init__(self):
        super(model_lstm, self).__init__()
        self.lstm0 = nn.LSTM(input_size=13 ,hidden_size=200, 
                             num_layers=1, batch_first=True)
        
        self.lstm1 = nn.LSTM(input_size=200 ,hidden_size=200, 
                             num_layers=1, batch_first=True)
        self.fc0   = nn.Linear(200, 2)
 
    def forward(self, x):
 
        out, hidden1 = self.lstm0(x) 
        out, _ = self.lstm1(out, hidden1) 
        out    = self.fc0(out) 
        return out   

model = model_lstm().to(device)
model

model_lstm(
  (lstm0): LSTM(13, 200, batch_first=True)
  (lstm1): LSTM(200, 200, batch_first=True)
  (fc0): Linear(in_features=200, out_features=2, bias=True)
)

3.训练模型

3.1定义训练函数

# 训练循环
def train(dataloader, model, loss_fn, optimizer):
    size = len(dataloader.dataset)  # 训练集的大小
    num_batches = len(dataloader)   # 批次数目, (size/batch_size，向上取整)

    train_loss, train_acc = 0, 0  # 初始化训练损失和正确率
    
    for X, y in dataloader:  # 获取图片及其标签
        X, y = X.to(device), y.to(device)
        
        # 计算预测误差
        pred = model(X)          # 网络输出
        loss = loss_fn(pred, y)  # 计算网络输出和真实值之间的差距，targets为真实值，计算二者差值即为损失
        
        # 反向传播
        optimizer.zero_grad()  # grad属性归零
        loss.backward()        # 反向传播
        optimizer.step()       # 每一步自动更新
        
        # 记录acc与loss
        train_acc  += (pred.argmax(1) == y).type(torch.float).sum().item()
        train_loss += loss.item()
            
    train_acc  /= size
    train_loss /= num_batches

    return train_acc, train_loss

3.2定义测试函数model

def test (dataloader, model, loss_fn):
    size        = len(dataloader.dataset)  # 测试集的大小
    num_batches = len(dataloader)          # 批次数目, (size/batch_size，向上取整)
    test_loss, test_acc = 0, 0
    
    # 当不进行训练时，停止梯度更新，节省计算内存消耗
    with torch.no_grad():
        for imgs, target in dataloader:
            imgs, target = imgs.to(device), target.to(device)
            
            # 计算loss
            target_pred = model(imgs)
            loss        = loss_fn(target_pred, target)
            
            test_loss += loss.item()
            test_acc  += (target_pred.argmax(1) == target).type(torch.float).sum().item()

    test_acc  /= size
    test_loss /= num_batches

    return test_acc, test_loss

3.3训练模型

loss_fn    = nn.CrossEntropyLoss() # 创建损失函数
learn_rate = 1e-4   # 学习率
opt        = torch.optim.Adam(model.parameters(),lr=learn_rate)
epochs     = 30

train_loss = []
train_acc  = []
test_loss  = []
test_acc   = []

for epoch in range(epochs):
    model.train()
    epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, opt)
 
    model.eval()
    epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)

    train_acc.append(epoch_train_acc)
    train_loss.append(epoch_train_loss)
    test_acc.append(epoch_test_acc)
    test_loss.append(epoch_test_loss)
    
    # 获取当前的学习率
    lr = opt.state_dict()['param_groups'][0]['lr']
    
    template = ('Epoch:{:2d}, Train_acc:{:.1f}%, Train_loss:{:.3f}, Test_acc:{:.1f}%, Test_loss:{:.3f}, Lr:{:.2E}')
    print(template.format(epoch+1, epoch_train_acc*100, epoch_train_loss, 
                          epoch_test_acc*100, epoch_test_loss, lr))
    
print("="*20, 'Done', "="*20)

Epoch: 1, Train_acc:56.2%, Train_loss:0.686, Test_acc:53.0%, Test_loss:0.697, Lr:1.00E-04
Epoch: 2, Train_acc:56.2%, Train_loss:0.683, Test_acc:53.0%, Test_loss:0.697, Lr:1.00E-04
Epoch: 3, Train_acc:56.2%, Train_loss:0.679, Test_acc:53.0%, Test_loss:0.695, Lr:1.00E-04
Epoch: 4, Train_acc:56.3%, Train_loss:0.675, Test_acc:53.0%, Test_loss:0.693, Lr:1.00E-04
Epoch: 5, Train_acc:56.8%, Train_loss:0.670, Test_acc:54.0%, Test_loss:0.690, Lr:1.00E-04
Epoch: 6, Train_acc:57.6%, Train_loss:0.665, Test_acc:55.4%, Test_loss:0.685, Lr:1.00E-04
Epoch: 7, Train_acc:59.5%, Train_loss:0.658, Test_acc:56.4%, Test_loss:0.679, Lr:1.00E-04
Epoch: 8, Train_acc:61.1%, Train_loss:0.649, Test_acc:58.4%, Test_loss:0.671, Lr:1.00E-04
Epoch: 9, Train_acc:64.2%, Train_loss:0.638, Test_acc:62.4%, Test_loss:0.660, Lr:1.00E-04
Epoch:10, Train_acc:67.7%, Train_loss:0.622, Test_acc:63.9%, Test_loss:0.645, Lr:1.00E-04
Epoch:11, Train_acc:69.2%, Train_loss:0.595, Test_acc:66.3%, Test_loss:0.617, Lr:1.00E-04
Epoch:12, Train_acc:72.6%, Train_loss:0.547, Test_acc:70.3%, Test_loss:0.590, Lr:1.00E-04
Epoch:13, Train_acc:74.3%, Train_loss:0.510, Test_acc:69.8%, Test_loss:0.569, Lr:1.00E-04
Epoch:14, Train_acc:75.0%, Train_loss:0.489, Test_acc:70.3%, Test_loss:0.558, Lr:1.00E-04
Epoch:15, Train_acc:75.7%, Train_loss:0.476, Test_acc:72.3%, Test_loss:0.550, Lr:1.00E-04
Epoch:16, Train_acc:76.2%, Train_loss:0.466, Test_acc:72.8%, Test_loss:0.544, Lr:1.00E-04
Epoch:17, Train_acc:77.4%, Train_loss:0.458, Test_acc:72.8%, Test_loss:0.538, Lr:1.00E-04
Epoch:18, Train_acc:77.5%, Train_loss:0.451, Test_acc:72.8%, Test_loss:0.532, Lr:1.00E-04
Epoch:19, Train_acc:77.4%, Train_loss:0.445, Test_acc:72.8%, Test_loss:0.527, Lr:1.00E-04
Epoch:20, Train_acc:78.1%, Train_loss:0.440, Test_acc:72.8%, Test_loss:0.523, Lr:1.00E-04
Epoch:21, Train_acc:78.2%, Train_loss:0.435, Test_acc:72.8%, Test_loss:0.520, Lr:1.00E-04
Epoch:22, Train_acc:78.6%, Train_loss:0.430, Test_acc:72.3%, Test_loss:0.517, Lr:1.00E-04
Epoch:23, Train_acc:79.1%, Train_loss:0.426, Test_acc:72.3%, Test_loss:0.514, Lr:1.00E-04
Epoch:24, Train_acc:79.5%, Train_loss:0.423, Test_acc:72.3%, Test_loss:0.512, Lr:1.00E-04
Epoch:25, Train_acc:79.4%, Train_loss:0.419, Test_acc:72.8%, Test_loss:0.511, Lr:1.00E-04
Epoch:26, Train_acc:79.5%, Train_loss:0.416, Test_acc:73.3%, Test_loss:0.509, Lr:1.00E-04
Epoch:27, Train_acc:79.9%, Train_loss:0.413, Test_acc:74.3%, Test_loss:0.508, Lr:1.00E-04
Epoch:28, Train_acc:80.2%, Train_loss:0.410, Test_acc:74.3%, Test_loss:0.507, Lr:1.00E-04
Epoch:29, Train_acc:80.5%, Train_loss:0.408, Test_acc:74.3%, Test_loss:0.507, Lr:1.00E-04
Epoch:30, Train_acc:80.6%, Train_loss:0.405, Test_acc:75.2%, Test_loss:0.507, Lr:1.00E-04
==================== Done ====================

4.模型评估

4.1Loss与acc

import matplotlib.pyplot as plt
#隐藏警告
import warnings
warnings.filterwarnings("ignore")               #忽略警告信息
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
plt.rcParams['figure.dpi']         = 100        #分辨率

from datetime import datetime
current_time = datetime.now() # 获取当前时间

epochs_range = range(epochs)

plt.figure(figsize=(12, 3))
plt.subplot(1, 2, 1)

plt.plot(epochs_range, train_acc, label='Training Accuracy')
plt.plot(epochs_range, test_acc, label='Test Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')
plt.xlabel(current_time) # 打卡请带上时间戳，否则代码截图无效

plt.subplot(1, 2, 2)
plt.plot(epochs_range, train_loss, label='Training Loss')
plt.plot(epochs_range, test_loss, label='Test Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()

​
在这里插入图片描述

​

相关文章：

  • UE 播放图像序列
  • 学习记录：初次学习使用transformers进行大模型微调
  • Docker镜像面试题及参考答案
  • 计算机毕业设计 ——jspssm513Springboot 的小区物业管理系统
  • HTML+CSS
  • 什么是数据治理？如何从数据治理中获得价值？
  • 【新人系列】Python 入门专栏合集
  • 【网络】TCP vs UDP详解( 含python代码实现)
  • AI如何通过大数据分析提升制造效率和决策智能化
  • hot100-栈 二分
  • 【我的 PWN 学习手札】IO_FILE 之 利用IO_validate_vtable劫持程序流
  • 【构建工具】Gradle 8中Android BuildConfig的变化与开启方法
  • WSL2下，向github进行push时出现timeout的问题
  • Web漏洞——命令注入漏洞学习
  • 【弹性计算】Guest OS
  • 内存资源分配
  • 视频推拉流EasyDSS直播点播平台授权激活码无效，报错400的原因是什么？
  • java后端开发day21--面向对象进阶（二）--继承进阶
  • Week 2 - Algorithm efficiency + Searching/Sorting
  • 浅谈HTTP及HTTPS协议
  • 李在明涉嫌违反《公职选举法》案将于15日进行首次重审公审
  • 特朗普宣布提名迈克·沃尔兹为下一任美国驻联合国大使
  • 李公明︱一周书记：数字文化的乌托邦精神与……算法时代的生存指南
  • 龚惠民已任江西省司法厅党组书记
  • 建设银行南昌分行引金融“活水”，精准灌溉乡村沃土
  • 蔡澜回应“入ICU观察”称未至于病危，助理：只是老毛病