数据挖掘-去长尾操作

def cut_col(data, col_name, cut_list):
    print('cutting', col_name)

    def _trans(array):
        count = array['box_counts']
        for box in cut_list:
            if count <= box:
                return 'count_' + str(box)
        return array[col_name]

    df_counts = pd.DataFrame(data[col_name].value_counts())
    df_counts.columns = ['box_counts']
    df_counts[col_name] = df_counts.index
    df = pd.merge(data, df_counts, on=col_name, how='left')
    column = df.apply(_trans, axis=1)
    return column
	
cut_col_dict = {
    ('pkgname', 'ver', 'reqrealip', 'adidmd5',
     'imeimd5', 'openudidmd5', 'macmd5', 'model', 'make'): [3],
    ('ip',): [3, 5, 10],
}

for cut_cols, cut_list in cut_col_dict.items():
    for col in cut_cols:
        data[col] = cut_col(data, col, cut_list)