Classification is the problem of identifying which set of categories based on observation features. The decision is based on a training set of data containing observations where category membership is known (supervised learning) or where category membership is unknown (unsupervised learning). A basic example is to determine the number from pixelated images in a built-in sklearn dataset. The script predicts 0-9 from the following images with a Support Vector Classifier.
In this case, the 8x8 (64) pixels are the input features to the classifier. The output of the classifier is a number from 0 to 9. The classifier is trained on 898 images and tested on the other 50% of the data. This is an example of supervised learning where the data is labeled with the correct number. An unsupervised learning method would not have the number labels on the training set. An unsupervised learning method creates categories instead of using labels.
See source code at https://apmonitor.com/do/index.php/Main/MachineLearningClassifier
