Predicting-the-Programming-Language-of-StackOverflow-Questions-using-Natural-Language-Processing

I used a LSTM based Deep learning model to predict the 16 most popular programming languages of 2019. I considered the title and body of a question to make prediction. I used a three-layer LSTM network as LSTMs are so effective for the sequential dataset and are widely preferred in NLP domain.

Data

I used StackSample dataset which is a collection of more than 1 million StackOverflow questions, answers and tags.

Requirements

Python 3.7

Keras 2.3

Numpy 1.18

Pandas 1.0

Matplotlib 3.1

NLTK 3.4.5

Re 2.2.1

Result

When tested on an unseen test set, I was able to achieve an accuracy of 82.34%.

Confusion Matrix

References

Predicting the Programming Language of Questions and Snippets of StackOverflow Using Natural Language Processing, Kamel Alrashedy