Classification of oscillometric envelope shape using frequent sequence mining

The shape of the oscillometric envelope is known to affect the accuracy of automatic noninvasive blood pressure (NIBP) measurement devices that use the oscillometric principle to determine systolic and diastolic blood pressures. This study proposes a novel shape classification method that uses data mining techniques to determine the characteristic sequences for different envelope shapes. The results indicate that the proposed method effectively determines the characteristic sequences for different subject groups. Subjects in the high- score group and in the low- score group tend to have an envelope with a broader plateau and are bell-shaped, respectively. This information about shape can be used for future determination of the correct algorithm for systolic and diastolic blood pressures determination in NIBP devices.