Paper Key : IRJ************614
Author: Shashank M
Date Published: 01 Jul 2024
Abstract
Despite a wealth of research on the link between indoor air pollution and the prevalence of allergic diseases, real-time big data and model predictability limitations prevent public health and environmental policies from developing predictive evidence for developing a preventive guideline for patients or vulnerable populations. Although the initiative is still in its early stages, the recent rise in popularity of IoT and machine learning techniques may offer supporting technologies for real-time big data collection and analysis for more precise prediction of allergy disease risks for evidence-based intervention. In this study, we provide a machine learningbased algorithm for predicting asthma risk. Using Internet- ofThings resources, the technology is totally deployed on a smartphone as a mobile health application. Peak Expiratory Flow Rates (PEFR), which are well-known asthma risk factors, are typically monitored using external tools such as peak flow meters. In this study, we use PEFR to identify a relationship between indoor particulate matter and the weather outside. The results of the PEFR are then divided into three risk categories, such as "Green'(Safe), "Yellow" (Moderate Risk), and "Red" (High Risk), and contrasted with the best peak flow number that was reached by everyone. The link between indoor PM levels and meteorological data and PEFR values is mapped using a convolutional neural network (CNN) architecture. The root mean square and mean absolute error accuracy measures of the suggested method are compared to those of deep neural network- based methods. In comparison to other methods addressed in the literature survey, the proposed method performs better according to these performance measures. On the smartphone, the entire configuration is implemented as an app. The input data is gathered using a Raspberry Pi-based IoT device. The cost-effectiveness of this aid in predicting the likelihood of asthma attacks can be demonstrated.
DOI Requested