更新 function.py

function.py

@@ -33,39 +33,43 @@ from sklearn.linear_model import LogisticRegression
 from sklearn.metrics import accuracy_score  
   
 def load_data():  
-    # 下载并加载病马数据集  
+    # 加载病马数据集  
     url = "http://archive.ics.uci.edu/ml/machine-learning-databases/horse-colic/horse-colic.data"  
     columns = [  
-        "surgery", "age", "hospital_number", "rectal_temp", "pulse", "respiratory_rate",  
+        "surgery", "age", "hospital_num", "rectal_temp", "pulse", "resp_rate",  
-        "temp_of_extremities", "peripheral_pulse", "mucous_membrane_color",  
+        "temp_of_extremities", "peripheral_pulse", "mucous_membrane", "capillary_refill_time",  
-        "capillary_refill_time", "pain", "peristalsis", "abdominal_distention",  
+        "muscle_tone", "abdominal_distension", "nasal_discharge", "vomit", "heart_rate",  
-        "nasal_discharge", "vomit", "heart_rate", "rectal_exam_fecal", "packed_cell_volume",  
+        "rectal_exam_fecal", "packed_cell_volume", "total_protein", "abdominocentesis_fluid",  
-        "total_protein", "abdominal_fluid_type", "abdominal_fluid_color", "outcome"  
+        "abdominocentesis_fluid_appearance", "leukocyte_count", "outcome", "surgical_summary",  
-    ] + ["unused_" + str(i) for i in range(1, 9)]  # 忽略最后8个无用特征  
+        "lesion_type", "lesion_site", "lesion_size"  
+    ]  
+    data = pd.read_csv(url, header=None, names=columns, na_values='?')  
   
-    data = pd.read_csv(url, header=None, names=columns, na_values="?")  
+    # 数据预处理  
-  
+    # 用0填充缺失值  
-    # 处理缺失值，用0填充  
     data.fillna(0, inplace=True)  
       
-    # 转换目标变量，将2（死亡）和3（安乐死）合并为0（死亡），1（存活）保持不变  
+    # 选择特征23作为目标变量，将取值2（死亡）和3（安乐死）合并为一个值0（死亡），并重新编码为二分类问题  
-    data["outcome"] = data["outcome"].apply(lambda x: 0 if x in [2, 3] else 1)  
+    data['outcome'] = data['outcome'].apply(lambda x: 0 if x in [2, 3] else 1)  
       
-    # 去掉特征3（hospital_number）和特征24~28（unused_1 到 unused_8）  
+    # 去掉特征3、特征24~28  
-    data.drop(columns=["hospital_number"] + ["unused_" + str(i) for i in range(1, 9)], inplace=True)  
+    data.drop(columns=['hospital_num', 'surgical_summary', 'lesion_type', 'lesion_site', 'lesion_size'], inplace=True)  
       
     # 特征和标签  
-    X = data.drop(columns=["outcome"])  
+    X = data.drop(columns=['outcome'])  
-    y = data["outcome"]  
+    y = data['outcome']  
       
-    # 划分训练集和测试集（这里简单按7:3划分）  
+    # 划分训练集和测试集（这里我们模拟划分，实际应使用给定的300个训练样本和68个测试样本）  
-    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)  
+    # 由于没有明确的划分方式，我们假设前300个为训练样本，后68个为测试样本（这通常不是实际做法，仅作为示例）  
+    # 注意：实际应用中应根据数据集提供的信息来准确划分训练集和测试集  
+    X_train, X_test = X.iloc[:300], X.iloc[300:]  
+    y_train, y_test = y.iloc[:300], y.iloc[300:]  
       
     return X_train, X_test, y_train, y_test  
   
 def train_logreg(X_train, y_train):  
-    # 训练逻辑回归模型  
+    # 训练对率回归模型  
-    model = LogisticRegression(max_iter=10000)  # 设置足够多的迭代次数以确保收敛  
+    model = LogisticRegression(max_iter=10000)  # 增加迭代次数以确保收敛  
     model.fit(X_train, y_train)  
     return model  
   
@@ -74,10 +78,15 @@ def predict_logreg(model, X_test):
     y_pred = model.predict(X_test)  
     return y_pred  
   
-# 示例调用  
+def evaluate_model(y_test, y_pred):  
+    # 评估模型性能  
+    accuracy = accuracy_score(y_test, y_pred)  
+    return accuracy  
+  
+# 主函数（用于测试和展示）  
 if __name__ == "__main__":  
     X_train, X_test, y_train, y_test = load_data()  
     model = train_logreg(X_train, y_train)  
     y_pred = predict_logreg(model, X_test)  
-    print(f"Accuracy: {accuracy_score(y_test, y_pred)}")
+    accuracy = evaluate_model(y_test, y_pred)  
-
+    print(f"Model Accuracy: {accuracy:.4f}")