Development of intelligent model for personalized guidance onwheelchair tilt and recline usage for people with spinal cord injury:Methodology and preliminary report

摘要

Wheelchair tilt and recline functions are two of themost desirable features for relieving seating pressure todecrease the risk of pressure ulcers. The effective guidance onwheelchair tilt and recline usage is therefore critical to pressureulcer prevention. The aim of this study was to demonstrate thefeasibility of using machine learning techniques to construct anintelligent model to provide personalized guidance to individualswith spinal cord injury (SCI). The motivation stems fromthe clinical evidence that the requirements of individuals varygreatly and that no universal guidance on tilt and recline usagecould possibly satisfy all individuals with SCI. We explored allaspects involved in constructing the intelligent model and proposedapproaches tailored to suit the characteristics of this preliminarystudy, such as modeling research participants, usingmachine learning techniques to construct the intelligent model,and evaluating the performance of the intelligent model. Wefurther improved the intelligent model’s prediction accuracy bydeveloping a twophase feature selection algorithm to identifyimportant attributes. Experimental results demonstrated thatour approaches showed promise they could effectively constructthe intelligent model, evaluate its performance, andrefine the participant model so that the intelligent model’s predictionaccuracy was significantly improved.