A study on air pollution index in Sabah and Sarawak using principal component analysis and artificial neural network/ Norwaziah Mahmud ... [et al.]

This study focuses on the identification of Sabah and Sarawak
air quality trends based on the data derived from the
Department of Environment (DOE). Five Malaysia’s monitoring
stations in Sabah and Sarawak were selected based on five
air pollutants for four years (2015-2018). This study aims to
classify the indicators of variable predictors using the Principal
Component Analysis (PCA) method and to compare the best
model to predict Air Pollution Index (API) in Sabah and
Sarawak using the Artificial Neural Network (ANN) model.
After running the varimax rotation, only two pollutants (PM10
and NO2) are the most significant pollutants out of the five
pollutants. These two pollutants were used as input layers in
Model B and the five pollutants were used as input layers in
Model A. These two models were used to compare the best
model in the ANN method. The output of ANN models was
evaluated through the coefficient of determination (R2
) and
Root Mean Square Error (RMSE). To identify the best model, the
highest value of R2 and the smallest value of RMSE were
declared. The findings indicate that the ANN technique has
been successfully implemented as a decision-making tool as
well as in solving problems for proper management of the
atmosphere.