In the context of machine learning, the term classification identifies an algorithmic procedure that assigns each new input datum (instance) to one of the possible categories (classes). If we consider only two classes, we talk about binary classification; otherwise we have a multi-class classification.

The classification falls into the supervised learning category, which permits us to classify new instances based on the so-called training set. The basic steps to follow to resolve a supervised classification problem are as follows:

  1. Build the training examples in order to represent the actual context and application on which to accomplish the classification.
  2. Choose the classifier and the corresponding algorithm implementation.
  3. Train the ...

Get Getting Started with TensorFlow now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.