generated from LufsX/Challenge-SVM
This commit is contained in:
+91
-58
@@ -30,95 +30,128 @@ name = "章程鸿"
|
||||
# 这里写自己的代码实现
|
||||
import numpy as np
|
||||
from sklearn.svm import SVC
|
||||
from sklearn.model_selection import train_test_split, KFold
|
||||
from sklearn.metrics import classification_report, accuracy_score
|
||||
from sklearn.preprocessing import StandardScaler
|
||||
|
||||
# 数据加载函数
|
||||
def load_data(filepath):
|
||||
"""
|
||||
从文件中加载数据，并进行必要的预处理，处理缺失值。
|
||||
从文件中加载数据，若不自行正确实现，则无法通过评测，**不记录成绩**。
|
||||
|
||||
Args:
|
||||
- filepath (str): 数据文件路径。
|
||||
- directory (str): 数据文件路径。
|
||||
|
||||
Returns:
|
||||
- X (numpy array): 特征数据。
|
||||
- y (numpy array): 标签数据。
|
||||
- np.array: 特征数据。
|
||||
- np.array: 标签数据。
|
||||
"""
|
||||
# 加载数据，假设分隔符是空格或者制表符
|
||||
data = np.genfromtxt(filepath, delimiter=' ', dtype=str)
|
||||
data = []
|
||||
labels = []
|
||||
with open(filepath, 'r') as file:
|
||||
for line in file:
|
||||
fields = line.strip().split()
|
||||
# 处理缺失值，用0填充
|
||||
fields = [0 if f == '?' else float(f) for f in fields]
|
||||
# 标签是倒数第6列
|
||||
label = int(fields[-6])
|
||||
# 特征为其他列（不包含倒数第6列标签）
|
||||
features = np.array(fields[: -6] + fields[-5:], dtype=float)
|
||||
# 标签映射：将 1->0, 2->1, 3->2
|
||||
if label == 1:
|
||||
label = 0
|
||||
elif label == 2:
|
||||
label = 1
|
||||
elif label == 3:
|
||||
label = 2
|
||||
else:
|
||||
print(f"无效标签，跳过此样本：{label}")
|
||||
continue # 跳过无效标签
|
||||
# 添加数据和标签
|
||||
data.append(features)
|
||||
labels.append(label)
|
||||
# 转换为numpy数组
|
||||
data = np.array(data)
|
||||
labels = np.array(labels)
|
||||
return data, labels
|
||||
|
||||
# 处理缺失值，将 '?' 替换为 np.nan
|
||||
data[data == '?'] = np.nan
|
||||
|
||||
# 将非数值字段转为浮动类型
|
||||
data = data.astype(np.float64)
|
||||
|
||||
# 选择特征和标签
|
||||
X = data[:, :-1] # 特征数据（假设最后一列是目标变量）
|
||||
y = data[:, -1] # 最后一列作为目标变量
|
||||
|
||||
# 不合并死亡与安乐死，直接使用原始标签
|
||||
# 1 = lived, 2 = died, 3 = euthanized
|
||||
|
||||
# 将缺失值填充为 0（或其他合适的值）
|
||||
X = np.nan_to_num(X, nan=0)
|
||||
|
||||
return X, y
|
||||
|
||||
def train_svm(X, y):
|
||||
def train_svm(X, y, other_params=None):
|
||||
"""
|
||||
使用 SVM 核方法进行训练。
|
||||
支持向量机训练函数。
|
||||
|
||||
Args:
|
||||
- X (numpy array): 特征数据。
|
||||
- y (numpy array): 标签数据。
|
||||
- other_params (str): 其它参数，可选设置。若需设置其它参数，必须设置其默认值，否则将无法通过自动评测，成绩为零。
|
||||
|
||||
Returns:
|
||||
- model (SVC): 训练得到的 SVM 模型。
|
||||
- scaler (StandardScaler): 用于后续预测的数据标准化器。
|
||||
- SVC: 训练好的SVM模型。
|
||||
"""
|
||||
# 数据标准化
|
||||
scaler = StandardScaler()
|
||||
X_scaled = scaler.fit_transform(X)
|
||||
model = SVC(kernel='rbf', C=1.0, gamma='scale', random_state=42)
|
||||
model.fit(X, y)
|
||||
return model
|
||||
|
||||
# 使用RBF核的SVM模型
|
||||
model = SVC(kernel='rbf', decision_function_shape='ovo') # 采用RBF核，多类分类使用'ovo'
|
||||
model.fit(X_scaled, y)
|
||||
|
||||
return model, scaler
|
||||
|
||||
def predict_svm(model, scaler, X):
|
||||
def predict_svm(inX, model):
|
||||
"""
|
||||
使用训练好的SVM模型进行预测。
|
||||
支持向量机预测函数。
|
||||
|
||||
Args:
|
||||
- model (SVC): 训练好的SVM模型。
|
||||
- scaler (StandardScaler): 标准化器，用于测试数据的标准化。
|
||||
- X (numpy array): 输入特征数据。
|
||||
- inX (numpy array): 输入待分类的样本。
|
||||
- model (SVC): 训练得到的支持向量机模型。
|
||||
|
||||
Returns:
|
||||
- int: 预测标签。
|
||||
- int: 预测的分类标签。
|
||||
"""
|
||||
X_scaled = scaler.transform([X]) # 将输入特征进行标准化
|
||||
return model.predict(X_scaled)[0]
|
||||
return model.predict(inX.reshape(1, -1))[0]
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
# 加载数据
|
||||
X_train, y_train = load_data("./data/horse-colic.data")
|
||||
X_test, y_test = load_data("./data/horse-colic.test")
|
||||
X_train, y_train = load_data("./horse-colic.data")
|
||||
X_test, y_test = load_data("./horse-colic.test")
|
||||
|
||||
# 训练SVM
|
||||
model, scaler = train_svm(X_train, y_train)
|
||||
# 划分数据集（这里划分出一部分作为最终测试集，剩下的用于交叉验证训练）
|
||||
X_full_train, X_final_test, y_full_train, y_final_test = train_test_split(X_train, y_train, test_size=0.2, random_state=42, stratify=y_train)
|
||||
|
||||
# 测试单个数据点
|
||||
y_pred = predict_svm(model, scaler, X_test[11])
|
||||
print(f"预测结果: {y_pred}, 实际结果: {y_test[11]}")
|
||||
# 数据标准化
|
||||
scaler = StandardScaler()
|
||||
X_full_train = scaler.fit_transform(X_full_train)
|
||||
X_final_test = scaler.transform(X_final_test)
|
||||
|
||||
# 测试整个测试集并评估准确率
|
||||
correct = 0
|
||||
for i in range(len(X_test)):
|
||||
y_pred = predict_svm(model, scaler, X_test[i])
|
||||
if y_pred == y_test[i]:
|
||||
correct += 1
|
||||
# 定义要尝试的C值列表
|
||||
C_values = [0.1, 1, 10]
|
||||
best_accuracy = 0
|
||||
best_model = None
|
||||
kf = KFold(n_splits=5, shuffle=True, random_state=42)
|
||||
for C in C_values:
|
||||
avg_accuracy = 0
|
||||
for train_index, val_index in kf.split(X_full_train):
|
||||
X_train_fold, X_val_fold = X_full_train[train_index], X_full_train[val_index]
|
||||
y_train_fold, y_val_fold = y_full_train[train_index], y_full_train[val_index]
|
||||
model = SVC(kernel='rbf', C=C, gamma='scale', random_state=42)
|
||||
model.fit(X_train_fold, y_train_fold)
|
||||
y_pred_fold = model.predict(X_val_fold)
|
||||
accuracy_fold = accuracy_score(y_val_fold, y_pred_fold)
|
||||
avg_accuracy += accuracy_fold
|
||||
avg_accuracy /= 5
|
||||
if avg_accuracy > best_accuracy:
|
||||
best_accuracy = avg_accuracy
|
||||
best_model = model
|
||||
|
||||
accuracy = correct / len(X_test) * 100
|
||||
print(f"测试集准确率: {accuracy}%")
|
||||
# 使用在交叉验证中选出的最佳模型在最终测试集上进行评估
|
||||
y_pred = best_model.predict(X_final_test)
|
||||
|
||||
# 输出评估结果
|
||||
print("\n分类报告:")
|
||||
print(classification_report(y_final_test, y_pred, target_names=["存活", "安乐死", "死亡"]))
|
||||
|
||||
# 输出准确率
|
||||
accuracy = accuracy_score(y_final_test, y_pred)
|
||||
print(f"测试集准确率: {accuracy * 100:.2f}%")
|
||||
|
||||
# 测试单个数据
|
||||
test_sample = X_final_test[11] # 选择第12个样本（从0开始计数）
|
||||
y_pred_single = predict_svm(test_sample, best_model)
|
||||
print(f"\n单个样本预测结果: {y_pred_single}, 实际结果: {y_final_test[11]}")
|
||||
Reference in New Issue
Block a user