Skip to main content
. 2019 Feb 1;21(2):137. doi: 10.3390/e21020137

Figure 1.

Figure 1

Graph representing the developed algorithm to predict degree of motor fluctuation severity. The algorithm consists of three stages: (A) hybrid feature extraction to extract incremental features that represent longitudinal changes in any two sensor-signal segments (referred to rounds); (B) A classifier to detect distribution of changes in UPDRS III scores between different rounds in a subject’s data; and (C) A decision-making system to assign an ACUPDRS III score (0 to 3) according to the UPDRS III change distribution and predict degree of motor fluctuation severity.