当前位置：首页 > wzjs >正文

深圳制作网站推荐亚马逊开店流程及费用

wzjs 2025/7/27 17:17:38

深圳制作网站推荐,亚马逊开店流程及费用,wordpress+缩略图+api,村网站建设计划书🍨 本文为🔗365天深度学习训练营中的学习记录博客🍖 原作者：K同学啊 👉 学习目标： ● 思考本案例是否还有进一步优化的空间 🏡 我的环境： ● 语言环境：Python3.12.4 ●…

👉 学习目标：
● 思考本案例是否还有进一步优化的空间

🏡 我的环境：

● 语言环境：Python3.12.4
● 编译器：Jupyter Lab
● 深度学习框架：pyTorch
● 数据地址：上周的数据

一、数据预处理

1、设置GPU

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='cuda')

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  \
0  18054421   0  38       1.25       2.99        1.07  0.64  5.31  83   83   
1  18054422   0  31       1.15       1.99        0.84  0.50  3.98  85   63   
2  18054423   0  27       1.29       2.21        0.69  0.60  4.19  73   61   
3  18054424   0  33       0.93       2.01        0.66  0.84  3.60  83   60   
4  18054425   0  36       1.17       2.83        0.83  0.73  4.83  85   67   高血压史   尿素氮     尿酸  肌酐  体重检查结果  是否糖尿病  
0     0  4.99  243.3  50       1      0  
1     0  4.72  391.0  47       1      0  
2     0  5.87  325.7  51       1      0  
3     0  2.40  203.2  40       2      0  
4     0  4.09  236.8  43       0      0

我们查看数据的shape：

DataFrame.shape

(1006, 16)

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

二、数据分析

1、数据分布分析

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='是否糖尿病', y=col, data=DataFrame, palette='Set2')  # 使用不同颜色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()

2、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)

三、LSTM模型

1、划分数据集

from sklearn.preprocessing import StandardScaler  
from sklearn.preprocessing import MinMaxScaler# '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)  
# 变成 LSTM 所需形状：(batch_size, seq_len=1, input_size=13)
train_X = train_X.unsqueeze(1)  # [643, 1, 13]
test_X = test_X.unsqueeze(1)    # [161, 1, 13]train_X.shape, train_y.shape

(torch.Size([804, 1, 13]), torch.Size([804]))

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

3、定义模型

我们在原来K同学的模型基础上，加入Attention机制：

class Attention(nn.Module):def __init__(self, hidden_dim):super(Attention, self).__init__()self.attn = nn.Linear(hidden_dim, 1)  # 输出为1个attention scoredef forward(self, lstm_output):  # lstm_output: (batch, seq_len, hidden_dim)attn_weights = self.attn(lstm_output)  # (batch, seq_len, 1)attn_weights = torch.softmax(attn_weights, dim=1)  # 时间维度归一化context = torch.sum(attn_weights * lstm_output, dim=1)  # (batch, hidden_dim)return contextclass 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.attn = Attention(hidden_dim=200)  # 添加 Attention 层self.fc0 = nn.Linear(200, 2)  # 输出维度根据你的任务设置为2类分类def forward(self, x):  out, hidden1 = self.lstm0(x)   out, _ = self.lstm1(out, hidden1)   attn_out = self.attn(out)  # (batch, hidden_dim)out = self.fc0(attn_out)  # 只使用 Attention 后的上下文向量进行分类return outmodel = model_lstm().to(device)  
model

model_lstm((lstm0): LSTM(13, 200, batch_first=True)(lstm1): LSTM(200, 200, batch_first=True)(attn): Attention((attn): Linear(in_features=200, out_features=1, bias=True))(fc0): Linear(in_features=200, out_features=2, bias=True)
)

四、训练模型

# 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_lossdef 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 loss_fn = nn.CrossEntropyLoss()  # 创建损失函数  
learn_rate = 1e-3   # 学习率  
opt = torch.optim.Adam(model.parameters(), lr=learn_rate)  
epochs = 50 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:66.8%, Train_loss:0.644, Test_acc:75.2%, Test_loss:0.589, Lr:1.00E-03
Epoch: 2, Train_acc:75.1%, Train_loss:0.517, Test_acc:75.7%, Test_loss:0.467, Lr:1.00E-03
Epoch: 3, Train_acc:77.4%, Train_loss:0.462, Test_acc:76.7%, Test_loss:0.482, Lr:1.00E-03
Epoch: 4, Train_acc:78.9%, Train_loss:0.432, Test_acc:78.7%, Test_loss:0.435, Lr:1.00E-03
Epoch: 5, Train_acc:79.1%, Train_loss:0.418, Test_acc:78.7%, Test_loss:0.443, Lr:1.00E-03
Epoch: 6, Train_acc:79.9%, Train_loss:0.405, Test_acc:80.2%, Test_loss:0.424, Lr:1.00E-03
Epoch: 7, Train_acc:79.6%, Train_loss:0.396, Test_acc:80.2%, Test_loss:0.383, Lr:1.00E-03
Epoch: 8, Train_acc:80.2%, Train_loss:0.389, Test_acc:78.2%, Test_loss:0.369, Lr:1.00E-03
Epoch: 9, Train_acc:80.8%, Train_loss:0.380, Test_acc:79.7%, Test_loss:0.383, Lr:1.00E-03
Epoch:10, Train_acc:81.0%, Train_loss:0.373, Test_acc:79.7%, Test_loss:0.425, Lr:1.00E-03
Epoch:11, Train_acc:81.3%, Train_loss:0.368, Test_acc:79.2%, Test_loss:0.360, Lr:1.00E-03
Epoch:12, Train_acc:82.6%, Train_loss:0.361, Test_acc:78.7%, Test_loss:0.363, Lr:1.00E-03
Epoch:13, Train_acc:81.5%, Train_loss:0.351, Test_acc:82.2%, Test_loss:0.369, Lr:1.00E-03
Epoch:14, Train_acc:82.2%, Train_loss:0.348, Test_acc:79.7%, Test_loss:0.373, Lr:1.00E-03
Epoch:15, Train_acc:82.1%, Train_loss:0.336, Test_acc:81.7%, Test_loss:0.398, Lr:1.00E-03
Epoch:16, Train_acc:82.8%, Train_loss:0.333, Test_acc:81.2%, Test_loss:0.323, Lr:1.00E-03
Epoch:17, Train_acc:83.0%, Train_loss:0.329, Test_acc:82.2%, Test_loss:0.350, Lr:1.00E-03
Epoch:18, Train_acc:83.0%, Train_loss:0.330, Test_acc:82.7%, Test_loss:0.362, Lr:1.00E-03
Epoch:19, Train_acc:84.6%, Train_loss:0.313, Test_acc:81.2%, Test_loss:0.404, Lr:1.00E-03
Epoch:20, Train_acc:85.3%, Train_loss:0.310, Test_acc:82.2%, Test_loss:0.401, Lr:1.00E-03
Epoch:21, Train_acc:84.6%, Train_loss:0.317, Test_acc:83.2%, Test_loss:0.399, Lr:1.00E-03
Epoch:22, Train_acc:85.3%, Train_loss:0.307, Test_acc:83.2%, Test_loss:0.403, Lr:1.00E-03
Epoch:23, Train_acc:86.4%, Train_loss:0.300, Test_acc:83.2%, Test_loss:0.396, Lr:1.00E-03
Epoch:24, Train_acc:85.8%, Train_loss:0.311, Test_acc:82.7%, Test_loss:0.481, Lr:1.00E-03
Epoch:25, Train_acc:86.4%, Train_loss:0.289, Test_acc:82.2%, Test_loss:0.419, Lr:1.00E-03
Epoch:26, Train_acc:85.9%, Train_loss:0.298, Test_acc:82.2%, Test_loss:0.374, Lr:1.00E-03
Epoch:27, Train_acc:86.2%, Train_loss:0.287, Test_acc:80.7%, Test_loss:0.429, Lr:1.00E-03
Epoch:28, Train_acc:87.1%, Train_loss:0.293, Test_acc:82.2%, Test_loss:0.470, Lr:1.00E-03
Epoch:29, Train_acc:87.4%, Train_loss:0.289, Test_acc:80.7%, Test_loss:0.470, Lr:1.00E-03
Epoch:30, Train_acc:87.1%, Train_loss:0.290, Test_acc:81.7%, Test_loss:0.421, Lr:1.00E-03
Epoch:31, Train_acc:87.3%, Train_loss:0.284, Test_acc:80.2%, Test_loss:0.470, Lr:1.00E-03
Epoch:32, Train_acc:87.6%, Train_loss:0.279, Test_acc:80.2%, Test_loss:0.429, Lr:1.00E-03
Epoch:33, Train_acc:86.7%, Train_loss:0.290, Test_acc:79.7%, Test_loss:0.385, Lr:1.00E-03
Epoch:34, Train_acc:88.2%, Train_loss:0.273, Test_acc:80.2%, Test_loss:0.445, Lr:1.00E-03
Epoch:35, Train_acc:87.9%, Train_loss:0.277, Test_acc:78.2%, Test_loss:0.421, Lr:1.00E-03
Epoch:36, Train_acc:88.3%, Train_loss:0.268, Test_acc:79.2%, Test_loss:0.455, Lr:1.00E-03
Epoch:37, Train_acc:87.7%, Train_loss:0.264, Test_acc:79.7%, Test_loss:0.470, Lr:1.00E-03
Epoch:38, Train_acc:88.7%, Train_loss:0.267, Test_acc:80.2%, Test_loss:0.465, Lr:1.00E-03
Epoch:39, Train_acc:88.4%, Train_loss:0.266, Test_acc:78.2%, Test_loss:0.456, Lr:1.00E-03
Epoch:40, Train_acc:88.7%, Train_loss:0.261, Test_acc:78.7%, Test_loss:0.412, Lr:1.00E-03
Epoch:41, Train_acc:87.8%, Train_loss:0.265, Test_acc:79.2%, Test_loss:0.671, Lr:1.00E-03
Epoch:42, Train_acc:89.1%, Train_loss:0.254, Test_acc:78.7%, Test_loss:0.381, Lr:1.00E-03
Epoch:43, Train_acc:88.7%, Train_loss:0.254, Test_acc:77.7%, Test_loss:0.401, Lr:1.00E-03
Epoch:44, Train_acc:88.7%, Train_loss:0.246, Test_acc:78.2%, Test_loss:0.670, Lr:1.00E-03
Epoch:45, Train_acc:89.6%, Train_loss:0.246, Test_acc:78.7%, Test_loss:0.415, Lr:1.00E-03
Epoch:46, Train_acc:89.3%, Train_loss:0.244, Test_acc:78.7%, Test_loss:0.452, Lr:1.00E-03
Epoch:47, Train_acc:89.1%, Train_loss:0.246, Test_acc:78.2%, Test_loss:0.488, Lr:1.00E-03
Epoch:48, Train_acc:89.3%, Train_loss:0.245, Test_acc:79.7%, Test_loss:0.397, Lr:1.00E-03
Epoch:49, Train_acc:89.4%, Train_loss:0.237, Test_acc:78.2%, Test_loss:0.412, Lr:1.00E-03
Epoch:50, Train_acc:89.4%, Train_loss:0.237, Test_acc:78.7%, Test_loss:0.451, Lr:1.00E-03
==================== Done ====================

五、模型评估

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()