Computer experiments for epidemiological studies : theory, methodology and models

Abstract

There are many types of computer-based models built for various purposes, for demonstration, exploration, prediction, and data analysis to name a few. In this work, we are interested in one particular type: to conduct epidemiological experiments with computer-based models. This is motivated by its potential in offering a less costly, safer, unquestionably ethical and possibly more powerful alternative to the real world scientific studies. There has been extensive research work on computational models that simulate real-life applications which can be used for exploring what-if scenarios in public health concerns, but they often lack a systematic approach in designing and conducting experiments that mirrors its real-life counterpart – that of a rigorous randomized, controlled experiment lauded as the gold standard by medical practitioners.