IoT-aware extreme machine learning for efficient health monitoring.

Urban Health monitoring systems significantly improve public well-being. Internet technologies have enabled a new dimension of healthcare technology, generating a large amount of data for informed decision-making. Many researchers used machine learning-based classification because of their physical feature extraction. However, the accuracy level was not improved. The Kernel Slice Regressed Recursive Extreme Learning Machine (KSRRELM) Technique is developed for efficient healthcare monitoring with higher accuracy. The KSRRELM technique is used to identify health conditions of human well-being, including physiological and psychiatric parameters. Healthcare information is collected from patients at different locations. Recursive Extreme Learning Machine is used to analyze patients’ physical and mental health conditions. Collected patient information is considered input. Data preprocessing is carried out to handle missing values using proximity projection pursuit and outlier detection. kernel slice inverse Regression is used to analyze patients’ physiological and psychiatric parameters to identify whether the patient’s condition is normal or abnormal. The analyzed results are sent to the output layer. Efficient healthcare monitoring is carried out with higher accuracy. Observed results attained accuracy 95%, precision of 94%, recall of 97% and F1-score of 96%. KSRRELM technique realizes a reduction in error rate of 5%, time of 40 ms, and space complexity of 275 MB compared to traditional methods.