Educational Path Recommender System using NLP


Recommendations to right posts can make the difference when receiving knowledge in a topic on sites like StackOverflow. StackOverflow gives users the ability to ask questions about topics of their interest then receive answers from the community. Providing the right recommendations can allow for a more straightforward educational path for the user and avoid repeated questions due to the lack of knowledge in the site's knowledge base. In this paper, we propose a hybrid recommendation approach that uses traditional methods and modern methods. We take advantage of a recent recommender model which takes advantage of a probabilistic model of collaborative filtering and a text content-based model. We also combine the modern recommender with another recommender model of our design that takes a traditional Content-based approach for providing direct recommendations of a given target post to provide an educational path for the user.