Deep learning for recommending subscription-limited documents

Documents recommendation for a commercial, subscription-based online platform is important due to the difficulty in navigation through a large volume and diversity of content available to clients. However, this is also a challenging task due to the number of new documents added every day and decreasing relevance of older contents. To solve this problem, we propose deep neural network architecture that combines autoencoder with multilayer perceptron in a hybrid recommender system. We train our model using real-world historical data from commercial platform using interactions to capture user similarity and categorical document features to predict the probability of a user-document interaction. Our experimental results demonstrate the effectiveness of the proposed architecture. We plan to release our model in a commercial online platform to support a personalized user experience.