SciReader - a personalized recommender for scientific literature

With the recent explosion in biomedical research, it has has become increasingly important and yet challenging to keep up with the relevant literature. SciReader is a personalized recommender system that specifically aims to help researchers and practitioners in the biomedical community parse through the large volume of literature and filter publications that may be relevant and of interest to them.

SciReader is a cloud based service that uses novel algorithms to classify and cluster published biomedical corpora using topic modeling (Latent Dirichlet Allocation). Researchers (our typical users) provide basic information like topics/keywords of interest and journal papers. Users can create multiple ‘libraries’ (for different projects) and upload papers of interest to it directly from pubmed. SciReader provides a seamless iinterface to search the pubmed database as well.

Personalized recommendations based on relevancy, recency, impact factor and sentiment analysis are created daily and weekly email digests of relevant publications in your field of research are sent directly to your inbox. ScIreader also monitors the twitter feed and tweets relevant to your field of interest are available in real time.

SciReader was initially developed at Jonathan Pritchard lab, Dept of Genetics, Stanford School of Medicine, and is now maintained and operated by the Stanford Center for Genomics and Personalized Medicine (SCGPM). Figure below shows user centric workflow.