Time series pattern recognition via SoftComputing

Abstract

In this paper we develop two methods that are able to analyze and recognize patterns in time series. The first model is based on analytic programming (AP), which belongs to softcomputing. AP is based as well as genetic programming on the set of functions, operators and so-called terminals, which are usually constants or independent variables. The second one uses an artificial neural network that is adapted by backpropagation. Artificial neural networks are suitable for pattern recognition in time series mainly because of learning only from examples. There is no need to add additional information that could bring more confusion than recognition effect. Neural networks are able to generalize and are resistant to noise. On the other hand, it is generally not possible to determine exactly what a neural network learned and it is also hard to estimate possible recognition error. They are ideal especially when we do not have any other description of the observed series. This paper also includes experimental results of time series pattern recognition carried out with both mentioned methods, which have proven their suitability for this type of problem solving.