在python中替换示例，并限制每个类的样本数

import pandas as pd
import numpy as np

data = pd.read_csv('example_dataset.csv')

# 替换示例
data.replace('old_value', 'new_value', inplace=True)

# 限制每个类的样本数
class_counts = data['class'].value_counts()
max_samples = 100  # 设置每个类的最大样本数
for class_label, count in class_counts.items():
    if count > max_samples:
        indices = data[data['class'] == class_label].index
        drop_indices = np.random.choice(indices, count - max_samples, replace=False)
        data.drop(drop_indices, inplace=True)

data.to_csv('processed_dataset.csv', index=False)