Application of Octonions in the Cough Sounds Classification

Artificial neural networks (ANN), as computational models inspired by biological neural networks, have become a standard tool in computer science for the classification of various digital patterns. They are used primarily in applications dealing with image processing, speech recognition, time series prediction, robotics and others [1]. Traditional artificial neural networks use real numbers for input data representation. The internal parameters of ANN, so-called weights, are also represented as real numbers. Several problems can be encountered during the process of training ANN: