Faculty Mentor(s)

Dr. Omofolakunmi Olagbemi, Computer Science

Document Type


Event Date



Machine learning (ML) is a powerful tool with vast applications in pattern-recognition and identification tasks. Our goal was to explore different applications of machine learning and develop a working understanding of the processes required for the effective application of ML models to problem-solving. Using SciKit-Learn for traditional ML models and TensorFlow for neural networks, existing techniques were explored for two major categories of ML tasks: Regression and Classification modeling. This knowledge was then applied in a biomedical engineering pilot research study (in collaboration with Dr. Brooke Odle, Engineering) analyzing manual patient-handling tasks using data from inertial measuring units (IMUs) and force plates. These tasks are linked to low-back pain and injury in caregivers. The use of IMUs in biomedical engineering enables flexible and mobile data collection both within and outside the laboratory. However, the force plates which are used for measuring the ground reaction forces (GRFs) are not as amenable to being transported for data collection outside the lab. Thus, our proof-of-concept study aims to develop and validate an artificial neural network (ANN) that estimates the ground reaction forces resulting from tasks performed by participants which simulate those that might be performed by a caregiver performing patient-handling tasks. Using data obtained from two subjects, a neural network was constructed and optimized. This model achieved a score of 0.9263 (92.63%), indicating that GRFs can be reasonably estimated with the use of an ANN. Future work would include expanding the study to involve more participants and include a wider variety of tasks, thereby improving the capacity of the ANN to generalize to fit more scenarios.


We acknowledge the support of the office of the Dean of Natural and Applied Sciences and the Computer Science department, both of Hope College, in funding this study.