A Machine Learning Approach for Running Grade Classification Using IMUs Data
摘要
This paper endeavors to advance Machine Learning techniques for biomechanical analysis through the classification of running grades. Employing data collected via three Inertial Measurement Unit (IMU) sensors, gait cycles were identified using a windowing technique, and a set of features from both time and frequency domain was identified and extracted. The methodology involved training and validating of three machine learning models, with performance assessed using standard metrics. Results revealed the superior performance of the quadratic support vector machine algorithm, boasting accuracy, recall, and precision metrics exceeding 99%.