An Iot Based Air Quality Monitoring And Air Pollutant Level Prediction System Using Machine Learning Approach – Dlmnn
Air pollution (AP) affects not only the environment but the body organs and respiratory system as well. Thus, a scientific air quality (AQ) monitoring in addition to the early system is needed for estimating the level of AP and also to precisely envisage the pollutant concentrations. The AQ prediction has considerably enhanced with the deployment of the sensor centered on the Internet of things (IoT). The entire existing techniques for AP predictions are costly and mostly have low accuracy. Thus, by employing Deep Learning Modified Neural Network (DLMNN) centered on IoT, a technique for monitoring and prediction is proposed here. The sensor values (SV) as of IoT devices are considered as input. After that, an attack detection mechanism is done on the received sensor data by utilizing the Gaussian kernel membership function (GMF) centered Adaptive Neuro-Fuzzy Inference System (ANFIS) termed as (GMF-ANFIS) to check whether the data is an attacked one or not. The system performs training to envisage the AP level. For training, the data as of the Air Quality Index (AQI) –UCI dataset is employed. In training, around ‘3’ operations are performed. Initially, data is preprocessed. Secondly, utilizing the Modified Dragonfly Optimization Algorithm (MDOA), the pertinent features as of the preprocessed data are chosen. Thirdly, the chosen features are classified into ‘6’ sorts of data by utilizing DLMNN. After training, testing is performed on the non-attacked SV. Finally, the results of testing are visualized. The experiment is implemented to analyze the proposed method’s performance. The outcomes exhibited that the DLMNN performs better when weighed against other existing algorithms for AQ prediction.