Literature Knowledge Graphs for Risk Modeling: a use case on Diabetes

Abstract

In this paper we present our preliminary efforts to support existing risk models with insights mined from sources of published literature such as PubMed. Information retrieval and Natural Language Processing (NLP) techniques were used to mine PubMed information and Knowledge Graphs (KGs) were subsequently created for a use case in type 2 diabetes. We used KGs to investigate the relevance of features in existing risk prediction models and highlight challenges and next steps.