Page 65 - THEi Student Applied Research Presentations 2024
P. 65

PREDICT IMPORTANT PARAMETERS OF
    GROUND REACTION FORCES AND RUNNING

    INJURY PATTERNS IN BIOMECHANICAL

    ANALYSIS USING MOTIONMETRIX AND
    MACHINE LEARNING



    ABSTRACT


    RESEARCH BACKGROUND
    Running, while highly beneficial to health, is typically linked to a significant risk of
    injury. Each year, a significant percentage of runners sustain injuries as a result of
    rigorous training programs and repetitive stress. However, there is new technology   MAK KA HO
    developments for more effective injury prevention and management strategies.
                                                                                 BSocSc (Hons) in Sports and
    METHODOLOGY                                                                  Recreation Management
    The  study  involved  65  marathon  runners,  with  the  project  aims,  protocols,  and   Department of Sport and Recreation
    procedures explained to them. Participants were invited to participate in Run Lap.
    Data were collected by questionnaire, MotionMetrix and clinical test. Data were then
    analysed and the Machine Learning Model was built.

    FINDINGS                                                                     OBJECTIVES

    The Random Forest Classifier identified 8 features related to injury patterns related to
    marathon runners. Initially, a set of four features related to ground reaction force was   The objectives  of  the research project
    evaluated: 'Vertical Force', 'Lateral Force', 'Contact Time', and 'Braking Force', across   are:
    three machine learning models. Subsequently, four additional features were included:   •  Identify  the  injury  patterns  which
    'Weekly Distance ', 'Hip Abduction Strength', 'Knee Alignment', and 'Knee-to-Wall   affect marathon runners
    Distance'. When predicting the cause of injury, the 4 ground reaction force factors has   •  To what extent can ground reaction
    lower accuracy than that of the other 4 features. The important features identified by   force (GRF) be used as a predictive
    the machine learning models are knee alignment, hip abduction strength, and weekly   parameter for running injuries via
    distance history.                                                                machine learning


    ABOUT THE INVESTIGATOR


    My name is MAK Ka Ho.
    My supervisor is Mr Indy HO Man Kit.






























 Student Applied Research Presentations 2024                                  Student Applied Research Presentations 2024  65
   60   61   62   63   64   65   66   67   68   69   70