fastText
Library for fast text classification and representation
...Nowadays, the dominant approach to build such classifiers is machine learning, that is learning classification rules from examples. In order to build such classifiers, we need labeled data, which consists of documents and their corresponding categories (or tags, or labels).