Python One Hot Encode多元变量

时间:2019-05-13 09:12:48

标签: python one-hot-encoding dummy-variable multivariate-partition

这是我的数据框:

Name   Job
A      Back-end Engineer
B      Front-end Engineer;Product Manager
C      Product Manager;Business Development;System Analyst

我想像这样将数据帧转换为虚拟(一种热编码):

Name   Back-end Engineer   Business Development   Front-end Engineer   Product Manager  System Analyst
A      1                   0                      0                    0                0
B      0                   0                      1                    1                0
C      0                   1                      0                    1                0

我尝试使用pandas.get_dummies,但是因为变量是多变量,所以失败了。

1 个答案:

答案 0 :(得分:1)

您可以尝试以下操作:

import pandas as pd
from collections import defaultdict


df = pd.read_csv("path/to/your.csv")

jobs = df["Job"]
job_list = set()
for job in jobs:
  job_names = job.split(";")
  for job_name in job_names:
    job_list.add(job_name)

new_df = defaultdict(list)
for index, row in df.iterrows():
  new_df["Name"].append(row["Name"])
  for job in job_list:
    if job in row["Job"]:
      new_df[job].append(1)
    else:
      new_df[job].append(0)

new_df = pd.DataFrame.from_dict(new_df)
new_df.to_csv("/path/to/new.csv")