val保存每个模型；LR统计请假m名单次数代码

LR_Process_concat_sheet.py

@@ -0,0 +1,16 @@
+import pandas as pd
+
+# 读取原始Excel文件的所有sheet
+excel_file = pd.ExcelFile('data/LeaveRecord.xlsx')
+df = pd.concat([excel_file.parse(sheet_name) for sheet_name in excel_file.sheet_names])
+
+# 统计每个学生的请假次数和姓名
+student_counts = df.groupby('学号').size().reset_index(name='请假次数')
+student_names = df.groupby('学号')['姓名'].unique().reset_index()
+student_counts['姓名'] = student_names['姓名'].apply(lambda x: ','.join(x))
+
+# 判断是否错误
+student_counts['是否错误'] = student_counts['姓名'].apply(lambda x: 1 if len(x.split(',')) > 1 else 0)
+
+# 保存结果到新的Excel文件
+student_counts.to_excel('output_LR_1.xlsx', index=False)

LR_Process_single_sheet.py

@@ -0,0 +1,19 @@
+import pandas as pd
+
+# 读取原始Excel文件
+df = pd.read_excel('data/LeaveRecord.xlsx', sheet_name=None)
+
+# 创建一个空的DataFrame来保存统计结果
+result_df = pd.DataFrame(columns=['学号', '请假次数'])
+
+# 遍历每个sheet
+for sheet_name, sheet_data in df.items():
+    # 统计每个学生的请假次数
+    student_counts = sheet_data['学号'].value_counts().reset_index()
+    student_counts.columns = ['学号', '请假次数']
+    
+    # 将统计结果添加到结果DataFrame中
+    result_df = pd.concat([result_df, student_counts], ignore_index=True)
+    
+# 保存结果DataFrame到新的Excel文件
+result_df.to_excel('output_LR.xlsx', index=False)

val.py

@@ -159,13 +159,14 @@ skf = StratifiedKFold(n_splits=n_splits, shuffle=True, random_state=random.randi
 input_size = src_features.shape[1]
 hidden_sizes = [32,128,32]
 output_size = len(set(src_labels)) # CrossEntropyLoss()损失函数类别数量就是输出size，sigmoid损失只适用于二分类，size为1，这里统一使用Cross
-lr = 0.0001 # learn rate 学习率
-num_epochs = 1000
+lr = 0.001 # learn rate 学习率
+num_epochs = 100
 batch_size = 128
 
 # 定义全局结果变量
 fold_accuracy=[]
 fold_loss=[]
+model_name=0
 
 # 遍历每个 fold
 for fold, (train_idx, test_idx) in enumerate(skf.split(src_features, src_labels)):
@@ -217,6 +218,9 @@ for fold, (train_idx, test_idx) in enumerate(skf.split(src_features, src_labels)
     fold_loss.append(test_loss)
     print(f'Accuracy for fold {fold}: {fold_accuracy[fold]*100} %, loss: {fold_loss[fold]}')
     print('--------------------------------')
+    # Save trained model
+    torch.save(model.state_dict(), 'psychology_model_val_{}.pth'.format(str(model_name)))
+    model_name+=1
 
 print('K-FOLD CROSS VALIDATION RESULTS')
 print(f'Fold accuracies: {fold_accuracy}')