A Novel Channel Minimization Technique for Diagnosis of Preterm Labor using Uterine Magnetomyography

The aim of this paper is to reduce the number of channels for Uterine Magnetomyography recording in order to facilitate its use in home uterine activity monitoring. We propose a technique of channel minimization using mutual information (CMMI) with the additional method of channel repeatability (CR). The result is a reduced channel set of 3 out of a large array of 148 with an improved classification accuracy of 61.67% from 35% using support vector machine (SVM) (kernel $=$ ‘rbf’) model.