Categorical data: Vocabulary and one-hot encoding

The term dimension is a synonym for the number of elements in a feature vector. Some categorical features are low dimensional. For example:

Feature name # of categories Sample categories
snowed_today 2 True, False
skill_level 3 Beginner, Practitioner, Expert
season 4 Winter, Spring, Summer, Autumn
day_of_week 7 Monday, Tuesday, Wednesday
planet 8 Mercury, Venus, Earth

When a categorical feature has a low number of possible categories, you can encode it as a vocabulary. With a vocabulary encoding, the model treats each possible categorical value as a separate feature. During training, the model learns different weights for each category.

For example, suppose you are creating a model to predict a car's price based, in part, on a categorical feature named car_color. Perhaps red cars are worth more than green cars. Since manufacturers offer a limited number of exterior colors, car_color is a low-dimensional categorical feature. The following illustration suggests a vocabulary (possible values) for car_color:

Figure 1. Each color in the palette is represented as a separate
      feature. That is, each color is a separate feature in the feature vector.
      For instance, 'Red' is a feature, 'Orange' is a separate feature,
      and so on.
Figure 1. A unique feature for each category.

Exercise: Check your intuition

True or False: A machine learning model can train directly on raw string values, like "Red" and "Black", without converting these values to numerical vectors.
True
During training, a model can only manipulate floating-point numbers. The string "Red" is not a floating-point number. You must convert strings like "Red" to floating-point numbers.
False
A machine learning model can only train on features with floating-point values, so you'll need to convert those strings to floating-point values before training.

Index numbers

Machine learning models can only manipulate floating-point numbers. Therefore, you must convert each string to a unique index number, as in the following illustration:

Figure 2. Each color is associated with a unique integer value. For
      example, 'Red' is associated with the integer 0, 'Orange' with the
      integer 1, and so on.
Figure 2. Indexed features.

Exercise: Check your intuition

Should your model train directly on the index numbers shown in Figure 2?
Yes
If the model trained on the index numbers, it would incorrectly treat each as a numerical value and consider "Black" (index number 5) to be 5 times more meaningful to the model than "Orange" (index number 1).
No
Your model shouldn't train on the index numbers. If it did, your model would treat each index number as a numerical value and consider "Black" (index number 5) to be 5 times more meaningful to the model than "Orange" (index number 1).

One-hot encoding

The next step in building a vocabulary is to convert each index number to its one-hot encoding. In a one-hot encoding:

  • Each category is represented by a vector (array) of N elements, where N is the number of categories. For example, if car_color has eight possible categories, then the one-hot vector representing will have eight elements.
  • Exactly one of the elements in a one-hot vector has the value 1.0; all the remaining elements have the value 0.0.

For example, the following table shows the one-hot encoding for each in car_color:

Feature Red Orange Blue Yellow Green Black Purple Brown
"Red" 1 0 0 0 0 0 0 0
"Orange" 0 1 0 0 0 0 0 0
"Blue" 0 0 1 0 0 0 0 0
"Yellow" 0 0 0 1 0 0 0 0
"Green" 0 0 0 0 1 0 0 0
"Black" 0 0 0 0 0 1 0 0
"Purple" 0 0 0 0 0 0 1 0
"Brown" 0 0 0 0 0 0 0 1

It is the one-hot vector, not the string or the index number, that gets passed to the feature vector. The model learns a separate weight for each element of the feature vector.

The following illustration suggests the various transformations in the vocabulary representation:

Figure 3. Diagram of the end-to-end process to map categories to
      feature vectors. In the diagram, the input features are 'Yellow',
      'Orange', 'Blue', and 'Blue' a second time.  The system uses a stored
      vocabulary ('Red' is 0, 'Orange' is 1, 'Blue' is 2, 'Yellow' is 3, and
      so on) to map the input value to an ID. Thus, the system maps 'Yellow',
      'Orange', 'Blue', and 'Blue' to 3, 1, 2, 2. The system then converts
      those values to a one-hot feature vector. For example, given a system
      with eight possible colors, 3 becomes 0, 0, 0, 1, 0, 0, 0, 0.
Figure 3. The end-to-end process to map categories to feature vectors.

Sparse representation

A feature whose values are predominately zero (or empty) is termed a sparse feature. Many categorical features, such as car_color, tend to be sparse features. Sparse representation means storing the position of the 1.0 in a sparse vector. For example, the one-hot vector for "Blue" is:

[0, 0, 1, 0, 0, 0, 0, 0]

Since the 1 is in position 2 (when starting the count at 0), the sparse representation for the preceding one-hot vector is:

2

Notice that the sparse representation consumes far less memory than the eight-element one-hot vector. Importantly, the model must train on the one-hot vector, not the sparse representation.

Outliers in categorical data

Like numerical data, categorical data also contains outliers. Suppose car_color contains not only the popular colors, but also some rarely used outlier colors, such as "Mauve" or "Avocado". Rather than giving each of these outlier colors a separate category, you can lump them into a single "catch-all" category called out-of-vocabulary (OOV). In other words, all the outlier colors are binned into a single outlier bucket. The system learns a single weight for that outlier bucket.

Encoding high-dimensional categorical features

Some categorical features have a high number of dimensions, such as those in the following table:

Feature name # of categories Sample categories
words_in_english ~500,000 "happy", "walking"
US_postal_codes ~42,000 "02114", "90301"
last_names_in_Germany ~850,000 "Schmidt", "Schneider"

When the number of categories is high, one-hot encoding is usually a bad choice. Embeddings, detailed in a separate Embeddings module, are usually a much better choice. Embeddings substantially reduce the number of dimensions, which benefits models in two important ways:

  • The model typically trains faster.
  • The built model typically infers predictions more quickly. That is, the model has lower latency.

Hashing (also called the hashing trick) is a less common way to reduce the number of dimensions.