A Medical Decision Aid Based on a Neural Network Model

Development of neural networks was one of the early topics of investigation in artificial intelligence in the 1950's. Research in this area lay dormant for a number of reasons, including the discovery that the single-layer perceptron-type networks could not represent some basic logical operations. Also, the state of computer hardware at that time could not accommodate neural networks of useful dimensions. In the 1980's interest in neural networks has increased enormously, due to advance in computer hardware and in theoretical developments which overcame shortcomings of the original perceptron model. In this paper, a neural network learning algorithm is described in the context of the development of a decision making aid. The model can accommodate a variety of data types. The approach is illustrated in a specific example in medical decision making.