Heuristic Based Optimized Classifier for HAR

Human Activity Recognition plays a pivotal role in various domains, from healthcare to wearable technology. This Paper presents a novel approach that combines heuristic-based feature selection with Particle Swarm Optimization to optimize the performance of HAR classifiers. We begin by employing heuristics to intelligently select relevant features from raw sensor data, reducing dimensionality and enhancing classification efficiency. Subsequently, PSO is applied to fine-tune classifier hyperparameters, achieving superior model accuracy and robustness. Our approach initiates by harnessing heuristics to autonomously discern and select pertinent features from raw sensor data. This intelligent feature selection not only trims the data dimensionality but also elevates the overall efficiency of the classification process. Subsequently, we apply PSO, a nature-inspired optimization algorithm, to meticulously fine-tune classifier hyperparameters.