Neural network based empirical models are used in a variety of data intensive medical applications. Physiological data acquired during hospital or outpatient visits presents complex waveform patterns that also contain significant amounts of noise. Neural network models can identify within those patterns specific signatures that are indicative of medical problems.

Other epidemiological phenomena are mined from the vast amount of data collected by organizations such as the Center for Disease Control and Prevention. The subtle interplay between environmental factors, disease symptoms, physician diagnoses, and individual genetic or behavioral characteristics can only be effectively analyzed and interpreted using models that learn from the data and from treatment outcomes.

Similarly, hospitals and HMOs use empirical models to monitor the effectiveness of treatment regimens, including physician performance and drug efficacy. When used in conjunction with national or even global data sources, empirical models can assist in identifying and tracking the spread of disease or can pinpoint the outbreak of specific health-related problems.

Finally, empirical models play important roles in the development and deployment of software and systems to analyze medical images. As the basic hardware and software to acquire and archive medical images continues to advance, empirical models increasingly are called upon to analyze images and generate diagnoses at rates that far exceed what a human practitioner could achieve.