Wave morphological classification of ECG-VCG's using a linear regression technique

Claus E. Liedtke, Naip Tuna

Research output: Contribution to journalReview articlepeer-review

Abstract

After a brief introduction to statistical decision theory, the paper describes the approximation of the probability p(x, i) of a measurement vector x and a diagnostic category i by a linear function in x using multiple multivariate regression techniques. The classification algorithm thus obtained is evaluated for increasing number of predictive measurements in terms of the percentage of correct classification for each diagnostic category, the overall percentage of correct classification and the sensitivity and specificity. For a fixed number of 16 ECG-VCG measurements the classification method is characterized by its misclassification matrix and its diagnostic operating curve. Based on a large test set of cases containing single and mixed diagnoses it could be shown that the algorithm gives very good results not only for single diagnostic cases but also in those situations where patients have several pathological conditions at a time.

Original languageEnglish (US)
Pages (from-to)189-194
Number of pages6
JournalKybernetes
Volume7
Issue number3
DOIs
StatePublished - Mar 1 1978

Fingerprint Dive into the research topics of 'Wave morphological classification of ECG-VCG's using a linear regression technique'. Together they form a unique fingerprint.

Cite this