The application :

This project is a flask application which allows to recommend Stack Overflow tags to any given message. It uses an hybrid unsupervised model built with both LDA and Word2Vec models in this project This models have been trained thanks to a querry built on https://data.stackexchange.com/stackoverflow/query/new . The corresponding query was :

      SELECT * FROM posts WHERE Id < 50000

Because of the time factor (computer science is changing rapidly), it is better to test the application on old messages. For example, testing it on messages providing from the querry:

      SELECT * FROM posts WHERE Id BETWEEN 50000 AND 50500

How to use it

First steps

Here are the instructions which allows to launch the application :

Once the application is launched

Just write (or copy/paste) the message of your choice and click on "GENERATE":

The recommendation results

The recommendation algorithm generates tags for you :

Acknowledgement :

This project was realized thanks to modifications on the HTML5 template provided by templated.co.