A Statistical Algorithm to Discover Knowledge in Medical Data Sources

Developing intelligent tools to extract information from data collections has long been of critical importance in fields such as knowledge discovery, information retrieval, pattern recognition, and databases. With the advent of electronic medical records and medical data repositories there is new potential to apply these techniques to the analysis of biomedical data sets. Looking for complex patterns within large biomedical data repositories and discovering previously unexpected associations can be of particular interest for understanding the physiology and functionality of the human body as well as tracing the roots of diseases. In the context of a research hospital these analyses may lead to further directed research, better diagnostic capabilities, and improved patient outcomes. This paper describes an implementation of a knowledge discovery algorithm aimed at such data sets.