A neural network expert system for computer-assisted analysis of blood-pressure data

The expert system described, based on a neural network architecture, analyzes a subject's anamnesis and 24-hour blood-pressure monitoring, to diagnose hypertension and give advice about choice, timing, and dosage of the most appropriate treatment. It also includes a set of explanation tools, such as the 'why?' and the 'what if . . .' functions, which emulate the ones that are typical of symbolic expert systems. Ongoing tests on 50 subjects rated both sensibility and specificity above 90%.<
>