An IoT-Based Asthma Intensity Prediction Using Classification Models

Asthma is a chronic disease of breathing that increases a patient's risk of an attack. So, taking care of the patients and supervising their health condition is essential to decrease the risk. This article centers on IoT, the ThingSpeak cloud platform, and machine learning methods employed for the prediction of asthma intensity. These sensors include the KY-038 audio sensor, SEN-11574 pulse sensor, and ESP-32 microcontroller, which collect live health data. Data transfer to ThingSpeak is secure. With data protocols, the devices transmit end-to-end data for storage and sharing. The combined use of these IoT sensors enables continuous health tracking for asthma. This data is visualized through a user-friendly mobile app built on platforms like MIT App Inventor. The app facilitates data exchange and even QR code generation, promoting high user engagement. The obtained data is then processed using machine learning algorithms, including Decision Tree Classifiers, Random Forest Classifiers, K Nearest Neighbors, and Logistic Regression. These commonly used algorithms help identify patterns within the data. By analyzing this data, the ML models can predict when an asthma attack is likely to occur and its potential intensity. This enables optimal clinical decision-making, allowing for proactive measures and efficacious asthma management for early intervention, such as using an inhaler or seeking medical attention.