Method conducted in a computer for classification of a time seri

Details Method conducted in a computer for classification of a time seri Method conducted in a computer for classification of a time seri

1998-09-03

2000-09-12

Hindenburg, Max

Surgery

Diagnostic testing

A61B 500

Patent

active

061170747

DESCRIPTION:

BRIEF SUMMARY
BACKGROUND OF THE INVENTION

1. Field of the Invention
The present invention is directed to a computerized method for classifying a time series containing a prescribable number of samples, such as an electrical signal.
2. Description of the Prior Art
In many technical fields it is of interest to draw conclusions about the future behavior of the time series from measured time series. The prediction of the future "behavior" of the time series ensues given the assumption that the time series comprises nonlinear correlations or linear correlations (statistical dependencies) between the samples of the time series.
This problem also obtains to considerable significance in various medical fields, for example in cardiology. Specifically in the problem area of sudden cardiac death, it can be vital to recognize early warning signs of sudden cardiac death in order to initiate counter-measures against the occurrence of sudden cardiac death as early as possible.
It is known that a time series of an electrocardiogram that is not correlated describes a heart that is not at risk with respect to sudden cardiac death. A heart at risk with respect to sudden cardiac death is described by a time series of the electrocardiogram that comprises non-linear correlations between the samples of the time series G. Morfill, "Komplexitatsanalyse in der Kardiologie," Physikalische Blatter, Vol. 50, No. 2, pp. 156-160, (1994). It is also known from the Morfill article to determine time series of an electrocardiogram that describe hearts that are at risk with respect to sudden cardiac death from the graphic phase space presentation of two successive heartbeats.
LICOX, GMS, Gesellschaft fur Medizinische Sondentecknik mbH, Advanced Tissue Monitoring discloses a method with which the time curve of the local oxygen voltage of the brain (tip02) can be determined.
The method disclosed in the Morfill article exhibits all of the disadvantages that are typical of empirical methods. In particular, the error susceptibility of graphic interpretations by a human, the problem of setting a threshold from which a time series is classified as at risk, as well as imprecisions in the presentation of the Fourier transform on the picture screen are considered disadvantageous in the known method.


SUMMARY OF THE INVENTION

An object of the present invention is a method for analytically classifying a time series with the assistance of a computer.
The above object is achieved in accordance with the principles of the present invention in a method for classifying a time series that contains a prescribable number of samples, such as an electrical signal, employing a computer, wherein a parameterized, dynamic set is obtained in the computer from the samples, the set describing non-linear correlations between the samples of the time series, and wherein the time series is classified in the computer on the basis on the parameterized, dynamic set.
In the inventive method, a parameterized, dynamic multiplicity is determined for a time series that contains a prescribable plurality of samples. Non-linear correlations between the samples of the time row are described with the parameterized, dynamic multiplicity on the basis of parameters that indicate the form of the parameterized, dynamic multiplicity. A classification of the time series is implemented on the basis of the parameterized, dynamic multiplicity.
Compared to the known, empirical inventive method, the method has the particular advantage that no error sources arise due to imprecise evaluation of the results. An unambiguous, replicatable classification of the time series is achieved by the analytic procedure. A further advantage is in the substantially increased speed with which the entire process of classification is implemented.
To allow the method to be conducted more rapidly, it is advantageous in the classification to classify the time series only into a first time series type and a second time series type. A first time series type thereby describes a time series wherein a non-linear correlation is found betw

REFERENCES:
patent: 5792062 (1998-08-01), Poon et al.
patent: 5938594 (1999-08-01), Poon et al.
"Komolexitatsanalyse in der Kardiologie," Morfill, Physikalische Blatter, vol. 50, No. 2, pp. 156-160, (1994).
Licox, GMS, Gesellschaft fur Medizinische Sondentechnik mbH, Advance Tissue Monitoring.
"Learning Time Series Evolution by Unsupervised Extraction of Correlations," Deco et al., Physical Revue E. vol. 51, No. 3, pp. 1780-1790 (Mar, 1995).
"Time Series Analysis of Complex Dynamics in Physiology and Medicine," Glass et al., Medical Progress Through Technology, vol. 19, No. 3, pp. 115-128 (Jan., 1993).

Affiliated with

Also associated with

Rating

Method conducted in a computer for classification of a time seri does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Method conducted in a computer for classification of a time seri, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method conducted in a computer for classification of a time seri will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-91330