Page 73 - THEi Student Applied Research Presentations 2024
P. 73

BIOMECHANICAL ANALYSIS IN LOWER LIMB
    ALIGNMENT IN PREDICTING RUNNING-RELATED

    INJURIES VIA MACHINE LEARNING


    ABSTRACT               Poster


    RESEARCH BACKGROUND

    The global popularity in running has resulted in an increase in running-related injuries.
    These injuries affect a significant portion of the running community and highlight the
    need for advanced biomechanical analysis in injury prevention and management.

    METHODOLOGY
    Employing MotionMetrix technology, the study collects detailed biomechanical data   YAM HO PONG AVERY
    on the lower limb alignment of the runners. Machine learning models—specifically
    RandomForestClassifier, XGBClassifier, and SVM—are applied to the collected data   BSocSc (Hons) in Sports and
    and identify patterns and predictors of injury risk. There were 59 participants in this   Recreation Management
    study, 44 males and 15 females aged 27-53, with at least 10km running experience.  Department of Sport and Recreation

    FINDINGS
    Model Findings through Training and Prediction Process: Among the machine learning
    models used, the XGBClassifier emerged as the most accurate, with a mean accuracy of
    0.6844 and precision of 0.7371; thus, XGBClassifier is effective in predicting accurate   OBJECTIVES
    true positives against false positives and negatives.
    Important Features and their Insights: Right-side hip abduction strength and right-side   Utilizing MotionMetrix technology for
    knee alignment were identified as significant features across all models, highlighting   non-invasive biomechanical data collection
    their pivotal role in predicting running-related injuries. The results proves that these   and machine learning for data analysis,
    biomechanical factors are crucial for understanding injury risk among runners.   this study seeks to identify key predictors
    Their significance likely stems from their impact on running mechanics and stress   of  injury  risk,  with  a  particular  focus  on
    distribution in the musculoskeletal system. Features that show a strong correlation   lower limb alignment.
    with the outcome across various scenarios within the training data tend to be weighted
    more heavily. For example, hip abduction strength and knee alignment might directly
    influence running posture, leading to a higher injury risk, hence their significance in
    the models.

    ABOUT THE INVESTIGATOR


    As a sports therapist, I would like to offer tailored rehabilitation services to athletes,
    focusing on their rapid recovery and return to peak performance. Ambitiously, I aspire
    to establish a sports therapy clinic in Hong Kong dedicated to providing top-tier
    services accessible to everyone, aiming to break barriers and ensure that high-quality
    sports therapy is available to individuals from all walks of life. My supervisor is Mr. HO
    Man Kit, Indy.





















 Student Applied Research Presentations 2024                                  Student Applied Research Presentations 2024  73
   68   69   70   71   72   73   74   75   76