Identification of Significant Features in Signals of Equine Cardiovascular System using ICA
Martin Kohut, Jiri Holcik, Jaroslav Hanak, Radovan Kabes
DOI:
Abstract
The paper deals with application of the Independent Component Analysis (ICA) for looking for causes of the sudden cardiovascular death in horses during surgery. The first chapter describes the solved problem and state of art. Next part describes the ICA method and its properties and an input data format. In the third and fourth parts results of the analysis are presented. It was found that not all the three important components contained in the input data must be also part of the output data set. The most frequent case (6 records of 10) is that only two components were sufficient for description of input data.