更新 function.py

function.py

@@ -26,69 +26,57 @@ name = "周佳成"
 
 
 # 这里写自己的代码实现
-import pandas as pd
+import pandas as pd  
-import requests
+import numpy as np  
-
+from sklearn.model_selection import train_test_split  
-# 下载病马数据
+from sklearn.linear_model import LogisticRegression  
-def download_horse_colic_data(url, filename):
+from sklearn.metrics import accuracy_score  
-    response = requests.get(url)
+  
-    with open(filename, 'wb') as f:
+def load_data():  
-        f.write(response.content)
+    # 下载并加载病马数据集  
-
+    url = "http://archive.ics.uci.edu/ml/machine-learning-databases/horse-colic/horse-colic.data"  
-# 处理并加载数据
+    columns = [  
-def load_and_preprocess_horse_colic_data(train_filename, test_filename):
+        "surgery", "age", "hospital_number", "rectal_temp", "pulse", "respiratory_rate",  
-    # 定义列名
+        "temp_of_extremities", "peripheral_pulse", "mucous_membrane_color",  
-    column_names = [
+        "capillary_refill_time", "pain", "peristalsis", "abdominal_distention",  
-        'surgery', 'age', 'hospital_number', 'rectal_temperature',
+        "nasal_discharge", "vomit", "heart_rate", "rectal_exam_fecal", "packed_cell_volume",  
-        'pulse', 'respiratory_rate', 'temperature_of_extremities', 
+        "total_protein", "abdominal_fluid_type", "abdominal_fluid_color", "outcome"  
-        'peripheral_pulse', 'mucous_membrane', 'capillary_refill_time',
+    ] + ["unused_" + str(i) for i in range(1, 9)]  # 忽略最后8个无用特征  
-        'pain', 'surgical_lesion', 'type_of_lesion', 'surgical_lesion_location',
+  
-        'outcome'  # 假设'outcome'是目标变量，1=lived, 2=died, 3=euthanized
+    data = pd.read_csv(url, header=None, names=columns, na_values="?")  
-    ]
+  
-
+    # 处理缺失值，用0填充  
-    # 读取训练数据
+    data.fillna(0, inplace=True)  
-    train_data = pd.read_csv(train_filename, delim_whitespace=True, header=None, names=column_names)
+  
-    # 读取测试数据
+    # 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  
-    test_data = pd.read_csv(test_filename, delim_whitespace=True, header=None, names=column_names)
+    data["outcome"] = data["outcome"].apply(lambda x: 0 if x in [2, 3] else 1)  
-
+  
-    # 处理缺失值：用0填充（这里假设除了'outcome'之外的所有特征都可能有缺失值）
+    # 去掉特征3（hospital_number）和特征24~28（unused_1 到 unused_8）  
-    for column in train_data.columns[:-1]:  # 排除目标变量列
+    data.drop(columns=["hospital_number"] + ["unused_" + str(i) for i in range(1, 9)], inplace=True)  
-        train_data[column].fillna(0, inplace=True)
+  
-    for column in test_data.columns[:-1]:  # 排除目标变量列
+    # 特征和标签  
-        test_data[column].fillna(0, inplace=True)
+    X = data.drop(columns=["outcome"])  
-
+    y = data["outcome"]  
-    # 处理目标变量'outcome'的缺失值
+  
-    train_data['outcome'].fillna(0, inplace=True)  # 可以选择适合的填充方法
+    # 划分训练集和测试集（这里简单按7:3划分，实际应用中应与原数据集划分保持一致）  
-    test_data['outcome'].fillna(0, inplace=True)
+    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)  
-
+  
-    # 将目标变量'outcome'转换为数值型
+    return X_train, X_test, y_train, y_test  
-    train_data['outcome'] = train_data['outcome'].astype(int)
+  
-    test_data['outcome'] = test_data['outcome'].astype(int)
+def train_logreg(X_train, y_train):  
-
+    # 训练对率回归模型  
-    return train_data, test_data
+    model = LogisticRegression(max_iter=10000)  # 设置足够多的迭代次数以确保收敛  
-
+    model.fit(X_train, y_train)  
-# 主函数
+    return model  
-def main():
+  
-    # 数据集URL
+def predict_logreg(model, X_test):  
-    train_url = 'http://archive.ics.uci.edu/ml/machine-learning-databases/horse-colic/horse-colic.data'
+    # 使用模型进行预测  
-    test_url = 'http://archive.ics.uci.edu/ml/machine-learning-databases/horse-colic/horse-colic-test.data'
+    y_pred = model.predict(X_test)  
-    
+    return y_pred  
-    # 保存数据集的文件名
+  
-    train_filename = 'horse_colic_train.data'
+# 示例调用（注：这些调用在自动评测时不会执行，仅供测试使用）  
-    test_filename = 'horse_colic_test.data'
+if __name__ == "__main__":  
-    
+    X_train, X_test, y_train, y_test = load_data()  
-    # 下载数据集
+    model = train_logreg(X_train, y_train)  
-    download_horse_colic_data(train_url, train_filename)
+    y_pred = predict_logreg(model, X_test)  
-    download_horse_colic_data(test_url, test_filename)
+    print(f"Accuracy: {accuracy_score(y_test, y_pred)}")
-    
-    # 加载并预处理数据
-    train_data, test_data = load_and_preprocess_horse_colic_data(train_filename, test_filename)
-    
-    # 打印前几行数据以验证
-    print("训练数据预览：")
-    print(train_data.head())
-    print("测试数据预览：")
-    print(test_data.head())
-
-if __name__ == "__main__":
-    main()