AMI prediction at presentation from ECG and clinical questionnaires

This paper describes a number of practical issues associated with the realisation of a neural-network-based decision aid for the early diagnosis of myocardial infarction or heart attack. We begin by outlining the problem domain and raising some key points that may affect the adoption, or otherwise, of a computer-based aid. We then focus briefly on the particular choice of a neural network to provide the underlying intelligence before moving on to consideration of the data gathering and clinical study phases. It is hoped that by doing this some general lessons can be learned from the progress of the project