Ordinal discrimination of ECG orthogonal features.

A rigorous mathematical analysis of Electrocardiogram (ECG) data is presented in this paper, as a decision support system for ECG interpretation. The mathematical analysis: provides a new insight to the ECG data, provides flexibility for future understanding of the heart system or the ECG, and avoids the need of measurements of parameters and the disagreement concerning the standard criteria of those parameters. The system enables compressed storage of the ECG data and an easy and accurate retrieval ability. An "ordinal straight line" method is presented for simple description of the ECG features as a combination of straight lines. A new developed discrimination method--the bisector method--is used for discrimination and classification between the normal and pathological groups. The physician's decision is based on both the physician's heuristic approach and the system's classification.