A prediction of electrocardiography signals by combining ARMA model with nonlinear analysis methods

Authors:
J. J. Águila;E. Arias;M. M. Artigao;J. J. Miralles
Affiliations:
University of Castilla-La Mancha, Albacete Research Institute of Informatics, Albacete, Spain;University of Castilla-La Mancha, Albacete Research Institute of Informatics, Albacete, Spain;University of Castilla-La Mancha, Applied Physics Dept., Albacete, Spain;University of Castilla-La Mancha, Applied Physics Dept., Albacete, Spain
Venue:
ACACOS'11 Proceedings of the 10th WSEAS international conference on Applied computer and applied computational science
Year:
2011

Citing 4
Cited 0

Extracting qualitative dynamics from experimental data

Physica D
Practical method for determining the minimum embedding dimension of a scalar time series

Physica D
Time Series Analysis, Forecasting and Control

Time Series Analysis, Forecasting and Control
Nonlinear Time Series Analysis

Nonlinear Time Series Analysis

Quantified Score

Hi-index	0.00

Visualization

Abstract

The main goal of the project A/6983/06, sponsored by the Spanish Agency for Latin America Collaboration, is analyzing and predicting electrocardiography signals of people which suffer post-traumatic stress. In order to make the prediction stage of this research, different techniques have been used. On the one hand, autoregressive moving average models which have been intensively used during decades for prediction purposes, and the other hand, nonlinear time series analysis methods which allow us to obtain some properties from measurement data records as for example the minimal embedding dimension. We performed a MATLAB implementation that combines autoregressive moving average model parametrized by the minimal embedding dimension. Results have been reported in terms of prediction error according to the best fit metric. This metric takes into account the horizon of predictability k. Thus, when considering k less than or equal to 4 instants of time, the best fit metric is above 80%.