Notice: This page requires JavaScript to function properly.
Please enable JavaScript in your browser settings or update your browser.OrdinalEncoder | Preprocessing Data with Scikit-learn
ML Introduction with scikit-learn
course content

Course Content

ML Introduction with scikit-learn

ML Introduction with scikit-learn

1. Machine Learning Concepts
What is MLTypes of Machine LearningTraining SetTypes of DataMachine Learning Workflow
2. Preprocessing Data with Scikit-learn
Scikit-learn ConceptsGetting Familiar with DatasetDealing with Missing ValuesImpute Missing ValuesOrdinalEncoderOne-Hot EncoderLabelEncoderEncode Categorical VariablesWhy Scale the Data?StandardScaler, MinMaxScaler, MaxAbsScalerScale the FeaturesPreprocessing Summary
3. Pipelines
What is PipelineColumnTransformerPipelineCreating a Pipeline ChallengeFinal EstimatorCreate a Complete ML Pipeline
4. Modeling
ModelsKNeighborsClassifierEvaluating a Model. Train-Test split.Cross-ValidationEvaluate the Model with Cross-ValidationGridSearchCVThe Flaw of GridSearchCVTune Hyperparameters with RandomizedSearchCVModeling SummaryPutting It All Together

bookOrdinalEncoder

The next problem we will solve is categorical data. Recall that there are two types of categorical data.

Ordinal data follows some natural order, while nominal does not.
Since there is a natural order, we can encode categories to the numbers in that order.
For example, we would encode the 'rate' column containing 'Terrible', 'Bad', 'OK', 'Good', and 'Great' values like:

  • 'Terrible' – 0;
  • 'Bad' – 1;
  • 'OK' – 2;
  • 'Good' – 3;
  • 'Great' – 4.

To encode ordinal data, OrdinalEncoder is used. It just encodes the categories to 0, 1, 2, ...

Here is an image showing how it works.

OrdinalEncoder is easy to use like any other transformer. The only difficulty is to specify the categories argument correctly.
Let's look at the example of use. We have a dataset (not the Penguins dataset) with the 'education' column. Let's look at its unique values.


              
12345

            
import pandas as pd

df = pd.read_csv('https://codefinity-content-media.s3.eu-west-1.amazonaws.com/a65bbc96-309e-4df9-a790-a1eb8c815a1c/adult_edu.csv')

print(df['education'].unique())
copy

We need to create a list of ordered categorical values, in this case, from 'HS-grad' to 'Doctorate'.


              
1234567891011121314

            
import pandas as pd
from sklearn.preprocessing import OrdinalEncoder

# Load the data and assign X, y variables
df = pd.read_csv('https://codefinity-content-media.s3.eu-west-1.amazonaws.com/a65bbc96-309e-4df9-a790-a1eb8c815a1c/adult_edu.csv')
y = df['income'] # 'income' is a target in this dataset
X = df.drop('income', axis=1)
# Create a list of categories so HS-grad is encoded as 0 and Doctorate as 6
edu_categories = ['HS-grad', 'Some-college', 'Assoc', 'Bachelors', 'Masters', 'Prof-school', 'Doctorate']
# Initialize an OrdinalEncoder instance with the correct categories
ord_enc = OrdinalEncoder(categories=[edu_categories])
# Transform the 'education' column and print it
X['education'] = ord_enc.fit_transform(X[['education']])
print(X['education'])
copy

Note

OrdinalEncoder is mostly used to transform the features (X variable). And the X variable usually is a DataFrame containing more than 1 column.
Because of that, the categories argument allows specifying the list of categories for each column, e.g., categories=[col1_categories, col2_categories].
And if you want to transform only 1 column, you should still pass a list containing another list, e.g., categories=[col1_categories].
That's also the reason the .fit_transform() method expects the DataFrame and doesn't work with Series, so you need to pass df[['column']] to transform only one column.

Everything was clear?

How can we improve it?

Thanks for your feedback!

Section 2. Chapter 5
some-alt