Automatic Classification of Source Code Archives by Programming Language: A Deep Learning Approach

摘要

This paper proposes the use of a Deep Learning technique, the Long Short-Term Memory (LSTM) recurrent neural network, for the automatic classification of source code archives by programming language. Experiments show that this simple recurrent neural network architecture gives promising results in accuracy compared to the Naive Bayes classifier, currently used by Linguist, one of the most popular programming language classifiers.