Graph Auto Encoders for Content-based Document Recommendation System

In the modern world,recommendation system have a good potential to improve business outcomes and to improve user experience with better accuracy combined with more generalisation to new unseen query.We propose a content-based document recommendation system which retrieves the most similar documents to the query given by the user.The embedded documents are fed into Graph Auto Encoder (GAE) and query through sparse coder to do latent representation which are finally used to give the user with perfect recommendations.The proposed method got evaluated with an average accuracy value of 93.91%.