Bayesian Expert

Our probabilistic logic Bayesian net service, BayesianExpert, makes the manual creation of Bayesian networks based on expert rules and statistics directly from the medical literature easy. Bayesian nets are capable of holistically combining many reductionist relations from medicine, and can also be used to learn new probabilities from data and judge anomalous data for verification purposes. Most importantly, Bayesian networks model causal relationships and further model epidemiological inference that can tease out confounding variables in observational data. 

To facilitate participation of the scientific community Rejuve has made a scientist-friendly rule based Bayesian Network available for modeling scientific theories and data, as one of the ways to contribute a model that can potentially become part of a model ensemble that discovers a longevity treatment which the modeler will be compensated for in tokens.

BayesianExpert uses logical rules such as any_of, all_of and avg to combine variables according to soft computation of probabilities. Special dependency rules, like relative_risk and sensitivity, facilitate modeling directly from meta-analyses and systematic reviews of clinical trials, with linear programming behind the scenes to complete the picture of the relationship of variables to each other. Our net also has an explanation module, which is used to find and convey to the user the one thing they can do that will most change their health state.