What is Classification?

Assigning data into categorical compartments (ex: cat/dog, apple/pear/banana, etc.) is Classification

More formally, Classification refers to the process of assigning an input vector of features into a pre-determined set of classes. 

Classification is an example of Supervised learning, where we learn from training data to make predictions.

For example:

binary classification

  • predicting if the Stock Market with go UP or DOWN tomorrow (two classes also known as Binary Classification). 
  • we might use economic indicators, company performance information, news sentiment etc, to guide our prediction

multi-class classification

  • classifying if a news article is about Sports, Politics, Entertainment, or international ( 4 classes ) 
  • we might use named entities, and other words mentioned in the text as input to our classification model

multi-label classification

  • Predicting the genres of a movie into Action, Adventure, Comedy, Horror, etc. A movie can have multiple genres. 
  • we might use plot summary as input text to assign multiple labels to a movie.