Fig. 4From: Improving activity recognition using a wearable barometric pressure sensor in mobility-impaired stroke patientsFuzzy inference example of a Sit-to-Stand transition. First the input variables are fuzzified using the membership functions. Then for each rule, the implication function (minimum) computes the minimum degree of membership (Y-axis) across each fuzzy variable and reports into the “Event Activity” variable. The degree of membership for each rule/input is denoted by the shaded area. The “Event Activity” variables (resulting from each rule) are then aggregated by merging the shaded area (the maximum function) creating a polygon shape. The defuzzification consists of computing the centroid value of this shapeBack to article page