At Sensor Diagnosis for Smart Healthcare: Probability or Conditional Probability Based Approach vs. k-Nearest Neighbour

Abstract

In order to implement IoT-based health-care for improved quality of life, we have to deal with sensor and communication technologies. In this article, the authors propose an approach to analyse real-time data streaming from a patient's surface body sensors, which are to be looked upon in a small sliding window frame. Time series analysis of data from the sensors is effective in reducing the round-trip delay between patient and the medical server. Two algorithms are for the sensor, and odd measures are proposed based on joint probability and joint conditional probability. The proposed algorithms are to be SQL compliant, as traces of at-sensor UDBMS alongside elementary capabilities supports databases with a meagre amount of SQL, which is evident in the literature.