Sustainability | Free Full-Text | A Novel One-Dimensional CNN with Exponential Adaptive Gradients for Air Pollution Index Prediction


Air air pollution is among the world’s most vital challenges. Predicting air air pollution is important for air high quality analysis, because it impacts public well being. The Air Air pollution Index (API) is a handy software to explain air high quality. Air air pollution predictions can present correct info on the long run air pollution scenario, successfully controlling air air pollution. Governments have expressed rising concern about air air pollution resulting from its world impact on human well being and sustainable development. This paper proposes a novel forecasting mannequin utilizing One-Dimensional Deep Convolutional Neural Community (1D-CNN) and Exponential Adaptive Gradients (EAG) optimization to foretell API for a particular location, Klang, a metropolis in Malaysia. The proposed 1D-CNN–EAG exponentially accumulates previous mannequin gradients to adaptively tune the training charge and converge in each convex and non-convex areas. We use hourly air air pollution information over three years (January 2012 to December 2014) for coaching. Parameter optimization and mannequin analysis was completed by a grid-search with k-folds cross-validation. Outcomes have confirmed that the proposed strategy achieves higher prediction accuracy than the benchmark fashions when it comes to Imply Absolute Error (MAE), Root Imply Sq. Error (RMSE), Imply Absolute Share Error (MAPE) and the Correlation Coefficient (R-Squared) with values of two.036, 2.354, 4.214 and 0.966, respectively, and time complexity.
View Full-Text

Show Figures

Source link


Please enter your comment!
Please enter your name here