Project: Estimating Physical Fatigue Using Wearable Sensor Data Developed a machine learning pipeline to estimate perceived physical fatigue (Borg RPE scores) during shoulder exercises using data collected from Inertial Measurement Units (IMU) and Electromyography (EMG) sensors. The dataset, sourced from 27 participants, included biomechanical data across multiple repetitions at varying resistance levels.
Trained and evaluated K-Nearest Neighbors and Decision Tree regression models using Scikit-learn, with a 70/30 train-test split to establish baseline performance.
Built a more robust pipeline by applying preprocessing and feature selection techniques to improve Decision Tree regressor accuracy.
Implemented and justified a new regression model (e.g., Random Forest or Support Vector Regressor) not covered in class, aiming for optimal performance.
Compared model performance using execution time and accuracy, and documented iterative experimentation and reasoning behind design choices.
Tools: Python, Scikit-learn, Pandas, Jupyter/Google Colab.