Data Visualization & Exploration

Medicine is complex. Correlations between diseases, medications, symptoms, lab data and genomics are of a complexity that cannot be fully comprehended by humans anymore. Machine learning methods are required that help mining these correlations. But a pure technological or algorithm-driven approach will not suffice. We need to get physicians and other domain experts on board, we need to gain their trust in the predictive models we develop. Elsevier Health Analytics has developed a first version of the Medical Knowledge Graph, which identifies correlations (ideally: causations) between diseases, and between diseases and treatments. On a dataset comprising 6 million patient lives we have calculated 2000+ models predicting the development of diseases. Every model adjusts for ~3000 covariates. Models are based on linear algorithms. This allows a graphical visualization of correlations that medical personnel can work with.

February 16, 2017

 

Related Webinars

February 16, 2017 4:00pm EST

Predictive Analysis has become a game-changing component for today’s forward-looking organizations seeking to gain a competitive advantage. This functionality has…

February 16, 2017 1:00pm EST

If a volcano erupts in Iceland, why is Hong Kong your first supply chain casualty? And how do you figure…

February 16, 2017 11:00am EST

Continuous streams of data are generated in every industry from sensors, IoT devices, business transactions, social media, network devices, clickstream…