generated from LufsX/Challenge-kNN
71 lines
2.0 KiB
Python
71 lines
2.0 KiB
Python
python
|
|
import numpy as np
|
|
|
|
"""
|
|
qin
|
|
"""
|
|
id = "202210423206"
|
|
name = "詹宇豪"
|
|
|
|
def classify0(inX, dataSet, labels, k=3):
|
|
"""
|
|
KNN 分类算法。
|
|
Args:
|
|
- inX (numpy array): 输入待分类的样本。
|
|
- dataSet (numpy array): 训练数据集。
|
|
- labels (list or numpy array): 训练集对应的标签。
|
|
- k (int): KNN 算法参数，默认值为3。
|
|
Returns:
|
|
- int: 预测的分类标签。
|
|
"""
|
|
# 使用广播机制计算欧氏距离，提高效率
|
|
distances = np.linalg.norm(dataSet - inX, axis=1)
|
|
|
|
# 按距离从小到大排序并获取对应的索引
|
|
sorted_indices = np.argsort(distances)
|
|
|
|
# 选择距离最近的k个样本的标签
|
|
top_k_labels = labels[sorted_indices[:k]]
|
|
|
|
# 统计k个样本中出现最多的类别
|
|
predicted_label = np.bincount(top_k_labels).argmax()
|
|
|
|
return predicted_label
|
|
|
|
def normalize_data(data):
|
|
"""
|
|
对数据进行归一化处理，将每个特征缩放到 [0, 1] 范围。
|
|
Args:
|
|
- data (numpy array): 待归一化的数据集。
|
|
Returns:
|
|
- numpy array: 归一化后的数据集。
|
|
"""
|
|
min_vals = data.min(axis=0) # 每一列的最小值
|
|
max_vals = data.max(axis=0) # 每一列的最大值
|
|
ranges = max_vals - min_vals # 每一列的取值范围
|
|
norm_data = (data - min_vals) / ranges # 归一化公式
|
|
return norm_data, ranges, min_vals
|
|
|
|
def load_data(file_path):
|
|
"""
|
|
加载 TSV 格式数据文件，并对特征数据进行归一化。
|
|
"""
|
|
data = np.loadtxt(file_path, delimiter="\t")
|
|
features = data[:, :-1]
|
|
labels = data[:, -1].astype(int)
|
|
|
|
# 数据归一化
|
|
norm_features, _, _ = normalize_data(features)
|
|
|
|
return norm_features, labels
|
|
|
|
|
|
if __name__ == "__main__":
|
|
# 加载训练数据
|
|
train_file = "datingTrainSet2.tsv"
|
|
train_data, train_labels = load_data(train_file)
|
|
|
|
# 测试 classify0 函数
|
|
test_sample = np.array([65915, 3.868359, 0.128078])
|
|
result = classify0(test_sample, train_data, train_labels, k=3)
|
|
print("Predicted label:", result) |