Authors:
Markus Nilsson
;
Peter Funk
and
Ning Xiong
Affiliation:
Mälardalen University, Sweden
Keyword(s):
Decision support, Case-Based Reasoning, Time series, Biomedical sequences, Classification, Discrete Wavelet Transformations, Clustering, Respiratory Sinus Arrhythmia.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Artificial Intelligence and Decision Support Systems
;
Case-Based Reasoning
;
Enterprise Information Systems
;
Pattern Recognition
;
Symbolic Systems
;
Theory and Methods
Abstract:
Clinicians do sometimes need help with diagnoses, or simply need reinsurance that they make the right decision. This could be provided to the clinician in the form of a decision support system. We have designed and implemented a decision support system for the classification of time series. The system is called HR3Modul and is designed to assist clinicians in the diagnosis of respiratory sinus arrhythmia. Two parallel streams of physiological time series are analysed for the classification task. Patterns are retrieved from one of the time series by the support of the other time series. These patterns are transformed with wavelets and matched for similarity by Case-Based Reasoning. Pre-classified patterns are stored and are used as knowledge in the system. The amount of patterns that have to be matched for similarity is reduced by a clustering technique. In this paper, we show that classification of physiological time series by wavelets is a viable option for clinical decision support.