Smart and interactive healthcare system based on speech recognition using soft margin formulation and kernel trick

Healthcare in the current day must be sophisticated and interactive. Several issues need to be addressed when creating a complete healthcare environment, including accurate diagnoses, inexpensive modeling, simple design, reliable data transfer, and enough storage capacity. This research article proposes an efficient approach focusing on speech recognition using soft margin formulation and kernel trick to provide a simple and easily accessible monitoring system to elderly and impaired people. The objective is to develop a low-cost speech recognition system that enables easy access to the deployed Internet of Things devices in the smart assisted living facilities (smart hospital or home) through distributive supervisory mode. The suggested work utilized a Banana Pi M3 and a BPI-AI-Voice (Microsemi) module to enable remote connectivity of home appliances through mobile phones. This framework is developed to help elderly/disabled people communicate with home appliances using voice commands. The suggested platform intervention strengthens the speech recognition procedure by including advanced Support Vector Machines principles such as soft margin formulation and kernel trick with a Globally Optimal Reparameterization Algorithm. The suggested approach is indeed a Machine learning—based technology that uses voice commands to manage smart medical devices with 96.09% accuracy. The nominated speech recognition technology is flexible enough to use with scalability and ensures the privacy of the deployed devices. Thus, the proposed work helps to integrate our scheme of medical facilities to assist the senior citizens and others who are sick (physically disabled) in a most effective way.