In our previous articles, we discussed equal-width discretization and equal-frequency discretization. We can also perform discretization or binning using custom bin values. This type of discretization is called custom discretization.

For example, let’s read the titanic dataset. The dataset has a column named age. Now, let’s say those aged 0 to 5 years should be labeled as toddlers. Those aged 5 to 18 years should be labeled as young. Those who are more than 18, but less than 60 years should be labeled as adults. And the rest should be labeled as seniors. In other words, we want to discretize the age column based on custom bin values. We can use the following Python code for that purpose:

import pandas

df = pandas.read_csv("titanic.csv")
print(df.head())

df["age_group"] = pandas.cut(x=df["age"], bins=[0, 5, 18, 60, 100], labels=["toddler", "young", "adult", "senior"])
print(df.head())

Here, we are using the pandas.cut() function for discretization and the bins parameter of the function indicates the custom bin values. We are also labeling the bins after discretization.

The output of the above program will be:

   survived  pclass     sex   age  ...  deck  embark_town  alive  alone
0         0       3    male  22.0  ...   NaN  Southampton     no  False
1         1       1  female  38.0  ...     C    Cherbourg    yes  False
2         1       3  female  26.0  ...   NaN  Southampton    yes   True
3         1       1  female  35.0  ...     C  Southampton    yes  False
4         0       3    male  35.0  ...   NaN  Southampton     no   True

[5 rows x 15 columns]
   survived  pclass     sex   age  ...  embark_town  alive  alone age_group
0         0       3    male  22.0  ...  Southampton     no  False     adult
1         1       1  female  38.0  ...    Cherbourg    yes  False     adult
2         1       3  female  26.0  ...  Southampton    yes   True     adult
3         1       1  female  35.0  ...  Southampton    yes  False     adult
4         0       3    male  35.0  ...  Southampton     no   True     adult

[5 rows x 16 columns]