Exploring Edge Computing Capabilities in IoT Devices for Machine Learning-Based Stuttering Prediction Models

Abstract

The research explores the potential of edge computing in enhancing machine learning-based stuttering prediction models within the Internet of Things (IoT) framework. The objective is to evaluate the feasibility and effectiveness of leveraging edge computing for stuttering prediction, aiming to improve accuracy and reduce latency by processing data closer to the source. The methodology involves examining various edge computing frameworks and algorithms suitable for implementing machine learning models on resource-constrained IoT devices, employing techniques like federated learning and model optimization. Results demonstrate that deploying machine learning models on edge devices significantly reduces latency and enhances real-time prediction capabilities compared to traditional cloud-based approaches. However, challenges such as limited computational resources and energy constraints of IoT devices are identified, necessitating efficient model architectures and optimization techniques. The implications highlight the potential benefits of edge computing in improving the accessibility and efficiency of stuttering prediction systems, particularly in remote or resource-constrained environments. Moreover, the study contributes to advancing the integration of machine learning and IoT technologies for healthcare applications, offering innovative solutions in speech disorder diagnosis and intervention. In conclusion, this research showcases the feasibility and effectiveness of utilizing edge computing capabilities in IoT devices for developing machine learning-based stuttering prediction models, with future research focusing on exploring edge-native machine learning algorithms and optimizing model deployment strategies for diverse IoT environments.