Comparison of rule-based expert system, neural networks and probability

Abstract

This article "Comparison of rule-based expert system, neural networks and probability", shows the different models to solve a case of expert systems, agencies, therefore we are two areas of artificial intelligence and statistics and probability in order to solve this kind of case that requires knowledge. Considering similar cases-disease-like symptoms for different ways to solve them. In the case of chances you have to create a set of rules that reflects the interpretation of specialist assessment area (this is handled by the knowledge engineer) by executing a graph AND / OR and leading to language inference engine has to solve them. Another is using neural networks where input patterns are the symptoms and diseases are outputs, this will have to enter cases for the learning process, taking as the type of neural network back propagation. As a last case we use the probabilistic method of Bayes' theorem (taking the case history and background of patients). We use these methods to solve cases that require a specialist, where each has its particularity to give solutions (either in the process of learning, inference or evaluation by history), among other things.