151. Air quality prediction on IoT real-time sensor using supervised machine learning.
- Author
-
Joseph, K. and Lakshmipathy, M.
- Subjects
- *
SUPERVISED learning , *AIR quality , *AIR pollution , *AIR pollutants , *REGRESSION trees , *INTERNET of things , *CHEST pain , *SESSION Initiation Protocol (Computer network protocol) , *MACHINE learning - Abstract
The negative impact of air pollution on human health is substantial. Chest discomfort, haemorrhage, asthma, emphysema, and other respiratory issues may be caused by both short-term and long-term exposure to air pollution. According to the World Health Organisation, air pollution occurs both inside and outdoors and is responsible for around 7 million deaths annually throughout the globe. High levels of air pollutants, beyond WHO standards, are inhaled by about 92% of the global population. Predictions of air quality techniques made so far are both expensive and wrong. Current methods for predicting air quality have data size limitations. In the past, researchers have attempted to measure air quality using methods like probability and statistics, but these approaches are notoriously difficult to forecast and hence, provide inaccurate results. Advancements in machine learning algorithm technology have the potential to be very fast. Consequently, air quality forecast needs improvement for a healthier atmosphere. Using supervised machine learning techniques, this study builds an AQI prediction model using past data. It then presents an inexpensive, fast, and accurate IoT system for air quality prediction using real-time sensors and new assessment and prediction features. The air quality predictions were made using decision tree and supervised machine learning methods. Decision tree methods come in two varieties: classification trees, which use the data's class as their prediction result, and regression trees, which use a real number as their prediction result. Classification and regression tree, or CART, is the name given to this kind of graph. Lastly, surveys urge the public to receive air quality alerts, especially in locations with limited data, and air quality forecasts are developed and validated with a 96 percent accuracy rate. The Meteorological Department may depend on the air quality prediction model that is based on machine learning for forecasting the status of the air and its quality. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF