替换pandas.dataframe中的低频分类值，同时忽略NaNs

import pandas as pd
import numpy as np

# 创建一个示例DataFrame
df = pd.DataFrame({'A': ['apple', 'banana', 'apple', 'orange', np.nan],
                   'B': ['cat', 'dog', 'dog', 'cat', 'dog']})

# 统计每个分类值的频率
value_counts = df['A'].value_counts()

# 设置低频阈值
threshold = 2

# 确定低频分类值
low_freq_values = value_counts[value_counts < threshold].index

# 创建替换字典
replace_dict = {value: 'Other' for value in low_freq_values}

# 替换低频分类值并忽略NaNs
df['A'] = df['A'].replace(replace_dict, regex=True)

print(df)

        A    B
0   apple  cat
1   Other  dog
2   apple  dog
3  orange  cat
4     NaN  dog