Machine Learning - Initialize Model - Classification

This article describes the modules in Azure Machine Learning Studio that support the creation of classification models. You can use these modules to build binary or multiclass classification models.

For more information about Machine Learning Studio modules, and to learn how to combine modules to create complete machine learning experiments, see Machine learning modules.

About classification

Classification is a machine learning method that uses data to determine the category, type, or class of an item or row of data. For example, you can use classification to:

  • Classify email filters as spam, junk, or good.
  • Determine whether a patient's lab sample is cancerous.
  • Categorize customers by their propensity to respond to a sales campaign.
  • Identify sentiment as positive or negative.

Classification tasks are frequently organized by whether a classification is binary (either A or B) or multiclass (multiple categories that can be predicted by using a single model).

Create a classification model

To create a classification model, or classifier, first, select an appropriate algorithm. Consider these factors:

  • How many classes or different outcomes do you want to predict?
  • What is the distribution of the data?
  • How much time can you allow for training?

Machine Learning Studio provides multiple classification algorithms. When you use the One-Vs-All algorithm, you can even apply a binary classifier to a multiclass problem.

After you choose an algorithm and set the parameters by using the modules in this section, train the model on labeled data. Classification is a supervised machine learning method. It always requires labeled training data.

When training is finished, you can evaluate and tune the model. When you're satisfied with the model, use the trained model for scoring with new data.

List of modules

The Classification category includes the following modules:

Examples

For examples of classification in action, see the Azure AI Gallery.

For help choosing an algorithm, see these articles:

See also