Learn My Sites is a smart website article aggregator that uses the websites you specify along with advanced machine learning algorithms to show you the best content for you on your favorite websites.

Our project utilizes word embeddings using Natural Language Processing and scrapes the provided news websites with AJAX in order to obtain these word vector representations of articles.

It then runs unsupervised clustering on these word vectors with K-means to obtain clusters of relevant articles.

The outputs are then a click serve of relevant articles grouped by cluster. In addition we run t-SNE for JS to obtain a 2D embedding of these 4,000+ dimensional word embedding to visually represent the cluster space of articles.

Share this project: