generated from LufsX/Experiment-kNN
This commit is contained in:
+143
-82
@@ -1,105 +1,166 @@
|
||||
# function.py
|
||||
import numpy as np
|
||||
import os
|
||||
from scipy.spatial.distance import mahalanobis
|
||||
from collections import Counter
|
||||
|
||||
# 学号和姓名（这部分在实际代码中可能不需要，但为了完整性保留）
|
||||
id = "202210423111"
|
||||
name = ""
|
||||
"""
|
||||
实验说明:
|
||||
- 自行编写实现数据加载与 kNN 的手写数字分类算法。
|
||||
- 函数输入: 待分类的样本 input_vector, 训练数据集 dataSet, 训练集对应的标签 labels。
|
||||
- 函数返回: 预测的分类标签。
|
||||
|
||||
注意事项:
|
||||
- 在 function.py 中实现 kNN 分类算法。
|
||||
- 可以自行导入其他库，需要在 requirements.txt 添加依赖，若未在 requirements.txt 中添加使用到的依赖，则无法自动评测，成绩为零。
|
||||
- 自动评测脚本仅会加载 classify 和 load_data 函数，请勿修改函数名。
|
||||
- 训练集和测试集需手动下载解压，并放入当前目录下。
|
||||
|
||||
测评环境:
|
||||
- CPU: Intel Xeon(R) CPU E3-1231 v3 @ 3.40GHz
|
||||
- OS: Ubuntu 22.04.4 LTS
|
||||
- Python: 3.10.12
|
||||
- numpy: 2.1.1
|
||||
"""
|
||||
|
||||
|
||||
"""
|
||||
请在这里填写学号和姓名，不填写即成绩无效。
|
||||
"""
|
||||
id = "202210423111"
|
||||
name = "李劲"
|
||||
|
||||
|
||||
"""
|
||||
参数说明:
|
||||
- input_vector (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，请**自行设置默认值**，评测时仅会使用其默认值，不修改默认值将**无法通过**评测，**成绩为零**。
|
||||
返回值:
|
||||
- int: 预测的分类标签。
|
||||
"""
|
||||
|
||||
|
||||
def classify(input_vector, dataSet, labels, mean_vector, inv_cov_matrix, k=1):
|
||||
def classify(input_vector, dataSet, labels, k=3): # 设置 k 的默认值为 3
|
||||
"""
|
||||
KNN 分类算法，使用马氏距离。
|
||||
|
||||
Args:
|
||||
- input_vector (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，选择距离最小的 k 个样本进行投票。
|
||||
|
||||
Returns:
|
||||
- int: 预测的分类标签。
|
||||
"""
|
||||
# 确保输入向量与训练数据集的维度一致
|
||||
input_vector = input_vector.reshape(1, -1)
|
||||
# 计算训练数据集的均值向量
|
||||
mean_vec = np.mean(dataSet, axis=0)
|
||||
|
||||
# 计算每个数据点到输入向量的马氏距离
|
||||
mahalanobis_distances = []
|
||||
for data_point in dataSet:
|
||||
dist = mahalanobis(data_point.reshape(1, -1), input_vector, inv_cov_matrix)
|
||||
mahalanobis_distances.append(dist)
|
||||
# 计算训练数据集的协方差矩阵
|
||||
cov_mat = np.cov(dataSet, rowvar=False)
|
||||
|
||||
# 将距离列表转换为 NumPy 数组，方便后续操作
|
||||
mahalanobis_distances = np.array(mahalanobis_distances)
|
||||
# 计算输入样本到所有训练样本的马氏距离
|
||||
# 注意：mahalanobis 函数要求输入二维数组，因此需要对 input_vector 进行 reshape
|
||||
# 同时，dataSet 也需要是二维数组，如果它是三维的（例如有多个特征图），则需要先展平
|
||||
inv_covmat = np.linalg.inv(cov_mat) # 计算协方差矩阵的逆
|
||||
distances = []
|
||||
for row in dataSet:
|
||||
dist = mahalanobis(input_vector.reshape(1, -1), row.reshape(1, -1), inv_covmat)
|
||||
distances.append(dist)
|
||||
|
||||
# 获取距离最近的 k 个样本的索引
|
||||
k_indices = np.argsort(mahalanobis_distances)[:k]
|
||||
# 选择距离最小的 k 个样本的索引
|
||||
k_indices = np.argsort(distances)[:k]
|
||||
|
||||
# 获取最近邻样本的标签
|
||||
# 选择这些样本对应的标签
|
||||
k_nearest_labels = [labels[i] for i in k_indices]
|
||||
|
||||
# 找出出现次数最多的标签
|
||||
most_common = Counter(k_nearest_labels).most_common(1)[0][0]
|
||||
|
||||
return most_common
|
||||
# 返回出现次数最多的标签
|
||||
from collections import Counter
|
||||
most_common = Counter(k_nearest_labels).most_common(1)
|
||||
return most_common[0][0]
|
||||
|
||||
def load_dataset(directory):
|
||||
"""
|
||||
从文件夹中加载数据，并计算均值向量和协方差矩阵。
|
||||
"""
|
||||
features = []
|
||||
labels = []
|
||||
|
||||
# 遍历文件夹中的所有文件
|
||||
for filename in os.listdir(directory):
|
||||
if filename.endswith(".txt"): # 假设文件是 .txt 格式
|
||||
filepath = os.path.join(directory, filename)
|
||||
|
||||
try:
|
||||
label = int(filename.split('_')[0])
|
||||
image_vector = img_to_vector(filepath)
|
||||
features.append(image_vector)
|
||||
labels.append(label)
|
||||
except Exception as e:
|
||||
print(f"Error processing file {filepath}: {e}")
|
||||
|
||||
features = np.array(features)
|
||||
labels = np.array(labels)
|
||||
从文件夹中加载数据，并返回特征数据和标签数据。
|
||||
这里假设文件夹中的每个文件都是一个二值图像，文件名包含了标签信息。
|
||||
|
||||
# 假设cov_matrix是原始协方差矩阵
|
||||
|
||||
regularization_term = 1e-5 * np.eye(cov_matrix.shape[0])
|
||||
|
||||
cov_matrix_regularized = cov_matrix + regularization_term
|
||||
|
||||
inv_cov_matrix = np.linalg.inv(cov_matrix_regularized)
|
||||
# 计算均值向量和协方差矩阵
|
||||
|
||||
|
||||
return features, labels, mean_vector, inv_cov_matrix
|
||||
Args:
|
||||
- directory (str): 数据文件夹路径。
|
||||
|
||||
Returns:
|
||||
- np.array: 特征数据。
|
||||
- np.array: 标签数据。
|
||||
"""
|
||||
data = []
|
||||
labels = []
|
||||
for filename in os.listdir(directory):
|
||||
if filename.endswith(".txt"): # 假设图像数据以文本形式存储，且文件名包含标签
|
||||
label = int(filename.split('_')[0]) # 从文件名中提取标签，这里假设文件名格式为 "label_xxx.txt"
|
||||
img_vector = img_to_vector(os.path.join(directory, filename))
|
||||
data.append(img_vector)
|
||||
labels.append(label)
|
||||
return np.array(data), np.array(labels)
|
||||
|
||||
def img_to_vector(filename):
|
||||
"""
|
||||
将二值图像转换为向量。
|
||||
将二值图像（以文本形式存储）转换为向量。
|
||||
这里假设图像数据是以空格分隔的 0 和 1 组成的字符串。
|
||||
|
||||
Args:
|
||||
- filename (str): 图像文件路径。
|
||||
|
||||
Returns:
|
||||
- np.array: 二值图像向量。
|
||||
"""
|
||||
try:
|
||||
with open(filename) as file:
|
||||
lines = file.readlines()
|
||||
image_vector = np.array([int(char) for line in lines for char in line.strip()])
|
||||
return image_vector
|
||||
except Exception as e:
|
||||
print(f"Error reading file {filename}: {e}")
|
||||
return np.zeros(1024) # 假设图像向量长度为1024
|
||||
|
||||
if __name__ == "__main__":
|
||||
# 测试
|
||||
train_data, train_labels, mean_vector, inv_cov_matrix = load_dataset("trainingDigits")
|
||||
|
||||
# 注意：这里的测试数据应该是一个二维数组，形状为 (1, 1024)
|
||||
# 假设 test_data 是一个已经格式化为字符串的1024位二值图像
|
||||
test_data_str = (
|
||||
"00000000000000110000000000000000"
|
||||
# ... (省略的测试数据，确保总长度为1024)
|
||||
)
|
||||
test_data = np.array([int(char) for char in test_data_str]).reshape(1, 1024)
|
||||
|
||||
predicted_label = classify(
|
||||
test_data,
|
||||
train_data,
|
||||
train_labels,
|
||||
mean_vector,
|
||||
inv_cov_matrix,
|
||||
)
|
||||
|
||||
with open(filename, 'r') as f:
|
||||
img_str = f.read().strip()
|
||||
img_vector = np.array([int(char) for char in img_str]).reshape(1, -1)
|
||||
return img_vector.flatten() # 展平为一维向量
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
# 以下代码可用于测试，可自行随意修改
|
||||
train_data, train_labels = load_dataset("trainingDigits")
|
||||
|
||||
test_data = (
|
||||
"00000000000000110000000000000000"
|
||||
"00000000000011111000000000000000"
|
||||
"00000000000011111100000000000000"
|
||||
"00000000001111111110000000000000"
|
||||
"00000000001111111110000000000000"
|
||||
"00000000001111111100000000000000"
|
||||
"00000000011111111000000000000000"
|
||||
"00000000111111111000000000000000"
|
||||
"00000000111111110000000000000000"
|
||||
"00000000111111110000000000000000"
|
||||
"00000000111111110000000000000000"
|
||||
"00000000111111110000000000000000"
|
||||
"00000000111111100000000000000000"
|
||||
"00000001111111100000000000000000"
|
||||
"00000001111111100000000000000000"
|
||||
"00000001111111100000000000000000"
|
||||
"00000001111111111111000000000000"
|
||||
"00000001111111111111111000000000"
|
||||
"00000001111111111111111100000000"
|
||||
"00000011111111111111111110000000"
|
||||
"00000011111111111111111110000000"
|
||||
"00000001111111111111111110000000"
|
||||
"00000001111111111111111110000000"
|
||||
"00000000111111111100111111100000"
|
||||
"00000000111111111000111111100000"
|
||||
"00000000011111111001111111100000"
|
||||
"00000000011111111111111111000000"
|
||||
"00000000001111111111111111000000"
|
||||
"00000000000111111111111111000000"
|
||||
"00000000000011111111111100000000"
|
||||
"00000000000001111111110000000000"
|
||||
"00000000000000000111100000000000"
|
||||
)
|
||||
predicted_label = classify(
|
||||
np.array([int(char) for char in test_data]).reshape(1, 1024),
|
||||
train_data,
|
||||
train_labels,
|
||||
)
|
||||
|
||||
print(f"预测的标签: {predicted_label}")
|
||||
Reference in New Issue
Block a user