将字符串列表转换为int列表以进行机器学习

Question

I have a table with a column corresponding to the education level : "phd", "undergrad" , etc.

I would like to change those features to 0, 1, 2.... . in order to use the data as an input of a machine learning algorithm.

Is there a way in Python to automatically map those string features to integers?

Answer 1

You can use enumerate if you want to have the same serial order.

lista = [ "phd", "undergrad", "etc"]
>>> [i for i in enumerate(lista)]
[(0, 'phd'), (1, 'undergrad'), (2, 'etc')]

Answer 2

It's called a dict (dictionary). Something like this:

edu_level = {
    "phd": 0;
    "master": 1;
    "undergrad":2, 
    ...
}

Look up how to work with dictionaries, perhaps using keys "Python dictionary tutorial".

Answer 3

You will likely not want to feed the data as it is to a machine learning algorithm - for instance, if a phd is a 2 and a masters is a 1, does that mean that the phd is twice as good? You might instead use "one hot encoding" and create a binary matrix (ie 1's & 0's) that you can feed.

There are multiple libraries to do this, ie one such: http://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.OneHotEncoder.html

Good luck!

  BA MS PHD Person A 0 0 0 Person B 1 0 0 Person C 0 0 1 ...