关键词: 结肠动力, 无创诊断, 支持向量机

Abstract: Due to the deficiency of clinic cases, character of patient and noise in the raw data, it is very difficult to diagnose noninvasive colonic motility complaints accurately. Support Vector Machine (SVM) is an efficient novel method originated from the statistical learning theory. It is powerful in machine learning to solve problems with finite samples. In this paper, SVM is employed in detecting colonic motility and the results are encouraged compared with conventional methods. The accuracy of nonlinear SVM classifier is especially high in all kinds of classifiers, which indicates the potential application of SVM in colonic motility diagnosis.

Key words: colonic motility, noninvasive diagnosis, support vector machine