Overlapping gait pattern recognition using regression learning for elderly patient monitoring.

Gait recognition in elderly patient monitoring is a standard process that employs medical healthcare systems, wearable sensors, motion capturing devices, and Information and Communication Technologies (ICT). The patterns of the patient movement are observed at different time instances for identifying the abnormality in gaits to provide better assistance. In this article, a novel Overlapping Gait Pattern Recognition method based on Regression Learning (RL) is introduced. This method classifies the gait pattern based on the direction of movement and angle of deviation of the patient at the initial stage. The analyses of differentiation are performed using RL for identifying the errors and differences in gait patterns through correlation. The errors are recurrently analyzed through different iterates for approximating the recognition accuracy in a reduced time. The classification of patterns through correlation and conditional analysis of the regression helps identify the errors through intense learning and deviation identification. The proposed method is found to achieve better recognition accuracy, fewer error rates, and smaller recognition delays for different gait patterns. [ABSTRACT FROM AUTHOR]