Fuzzy classification of hemodynamic trends and artifacts: experiments with the heart rate

Fuzzy set theory allows one to map inexact data, concepts, and events to fuzzy sets via user-defined membership functions. This paper describes a method for (1) robustly estimating the mean and slope of an arbitrary number of data points, (2) developing a set of fuzzy membership functions to classify various properties of heart rate trends, and (3) finding the longest consecutive sequence of heart rate data that fit a particular fuzzy membership function. Preliminary results indicate that fuzzy set theory has significant potential in the development of a clinically robust method for classifying heart rate data, trends, and artifacts.