2015_Pattern Recognition and Source Apportionment Modeling of Air Quality In Peninsular Malaysia Using Chemometric Methods

Bibliographic Details
Format: General Document
_version_ 1860797983572885504
building INTELEK Repository
collection Online Access
collectionurl https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3
copyright Copyright©PWB2025
country Malaysia
date 2015-09-01 15:20
format General Document
id 15271
institution UniSZA
originalfilename PATTERN RECOGNITION AND SOURCE APPORTIONMENT MODELING OF AIR QUALITY IN PENINSULAR MALAYSIA USING CHEMOMETRIC METHODS
person PDFsam Basic v4.2.10
Isiyaka Hamza Ahmad
recordtype oai_dc
resourceurl https://intelek.unisza.edu.my/intelek/pages/view.php?ref=15271
sourcemedia Server storage
Scanned document
spelling 15271 https://intelek.unisza.edu.my/intelek/pages/view.php?ref=15271 https://intelek.unisza.edu.my/intelek/pages/search.php?search=!collection3 General Document Malaysia Library Staff (Top Management) Library Staff (Management) Library Staff (Support) Terengganu English UniSZA East Coast Environmental Research Institute application/pdf 1.5 PDFsam Basic v4.2.10 Chemometrics Server storage Scanned document Universiti Sultan Zainal Abidin UniSZA Private Access Universiti Sultan Zainal Abidin SAMBox 2.3.4; modified using iTextSharp™ 5.5.10 ©2000-2016 iText Group NV (AGPL-version) 2015-09-01 15:20 139 PATTERN RECOGNITION AND SOURCE APPORTIONMENT MODELING OF AIR QUALITY IN PENINSULAR MALAYSIA USING CHEMOMETRIC METHODS 2015_Pattern Recognition and Source Apportionment Modeling of Air Quality In Peninsular Malaysia Using Chemometric Methods Copyright©PWB2025 Isiyaka Hamza Ahmad Pattern Recognition Source Apportionment Modeling Air Quality Peninsular Malaysia Chemometric Methods Pollution—Measurement Multivariate analysis Environmental chemistry—Statistical methods This study was carried out to model and understand the pattern and dynamic characteristic of atmospheric air pollution in two strategic regions (Klang Valley and Terengganu). Chemometric techniques such as hierarchical agglomerative cluster analysis (HACA), discriminant analysis (DA), principal component analysis (PCA), multiple linear regression (MLR) and artificial neural network (ANN) were applied on the five years datasets (2007-2011). HACA successfully classified the five monitoring sites into three clusters. DA gave a correct assignation of 82.37% (p˂0.05) indicating that all the seven parameters discriminate spatially. PCA was used to identify the major possible sources of pollution and attribute their origin to anthropogenic activities (automobiles, industries, biomass burning, power plants, aircraft and construction sites) and natural processes (photochemical oxidation of precursors). The percentage contribution of individual pollutants and source apportionment were modeled using a combined PCA-MLR. Based on the standardized coefficient, coefficient of determination and leave-one-out cross-validation method, PM10 appears to be the most significant parameter in cluster 1, while O3 strongly affect the air pollution index (API) in cluster 2 and 3. The result of the source category apportionment shows that the API value is influenced by pollutant gasses, non-gas pollutants, secondary pollutants and weather condition. ANN was used to develop three input combination models (model A, B and C) to predict the API value. Model A (raw data) gives the best prediction capability with R2 = 0.93 and RMSE = 4.87 compared with model B (principal components) and C (principal component scores). Furthermore, two receptor models (ANN and MLR) were compared in order to select the best model for predicting API at a high precision. It is statistically proven that the ANN has the ability to predict API with R2 = 0.93 and RMSE = 4.87 compared with MLR model where the R2 = 0.70, RMSE= 9.57 respectively. This study justifies the ability of chemometric technique to model the dynamic characteristics of air pollution, reduce the cost and time of spatial air quality monitoring program, understand the sources of pollutants and the most significant parameters affecting the value of API. These findings can be used by government and all concerned agencies for decision-making, problem-solving and environmental management. Water quality—Mathematical models Water quality—Measurement Environmental monitoring—Statistical methods Hydrology—Statistical methods Water—Pollution—Analysis Geochemical modeling Dissertations, Academic Thesis
spellingShingle 2015_Pattern Recognition and Source Apportionment Modeling of Air Quality In Peninsular Malaysia Using Chemometric Methods
state Terengganu
subject Chemometrics
Pollution—Measurement
Multivariate analysis
Environmental chemistry—Statistical methods
Water quality—Mathematical models
Water quality—Measurement
Environmental monitoring—Statistical methods
Hydrology—Statistical methods
Water—Pollution—Analysis
Geochemical modeling
Dissertations, Academic
summary This study was carried out to model and understand the pattern and dynamic characteristic of atmospheric air pollution in two strategic regions (Klang Valley and Terengganu). Chemometric techniques such as hierarchical agglomerative cluster analysis (HACA), discriminant analysis (DA), principal component analysis (PCA), multiple linear regression (MLR) and artificial neural network (ANN) were applied on the five years datasets (2007-2011). HACA successfully classified the five monitoring sites into three clusters. DA gave a correct assignation of 82.37% (p˂0.05) indicating that all the seven parameters discriminate spatially. PCA was used to identify the major possible sources of pollution and attribute their origin to anthropogenic activities (automobiles, industries, biomass burning, power plants, aircraft and construction sites) and natural processes (photochemical oxidation of precursors). The percentage contribution of individual pollutants and source apportionment were modeled using a combined PCA-MLR. Based on the standardized coefficient, coefficient of determination and leave-one-out cross-validation method, PM10 appears to be the most significant parameter in cluster 1, while O3 strongly affect the air pollution index (API) in cluster 2 and 3. The result of the source category apportionment shows that the API value is influenced by pollutant gasses, non-gas pollutants, secondary pollutants and weather condition. ANN was used to develop three input combination models (model A, B and C) to predict the API value. Model A (raw data) gives the best prediction capability with R2 = 0.93 and RMSE = 4.87 compared with model B (principal components) and C (principal component scores). Furthermore, two receptor models (ANN and MLR) were compared in order to select the best model for predicting API at a high precision. It is statistically proven that the ANN has the ability to predict API with R2 = 0.93 and RMSE = 4.87 compared with MLR model where the R2 = 0.70, RMSE= 9.57 respectively. This study justifies the ability of chemometric technique to model the dynamic characteristics of air pollution, reduce the cost and time of spatial air quality monitoring program, understand the sources of pollutants and the most significant parameters affecting the value of API. These findings can be used by government and all concerned agencies for decision-making, problem-solving and environmental management.
title 2015_Pattern Recognition and Source Apportionment Modeling of Air Quality In Peninsular Malaysia Using Chemometric Methods
title_full 2015_Pattern Recognition and Source Apportionment Modeling of Air Quality In Peninsular Malaysia Using Chemometric Methods
title_fullStr 2015_Pattern Recognition and Source Apportionment Modeling of Air Quality In Peninsular Malaysia Using Chemometric Methods
title_full_unstemmed 2015_Pattern Recognition and Source Apportionment Modeling of Air Quality In Peninsular Malaysia Using Chemometric Methods
title_short 2015_Pattern Recognition and Source Apportionment Modeling of Air Quality In Peninsular Malaysia Using Chemometric Methods
title_sort 2015_pattern recognition and source apportionment modeling of air quality in peninsular malaysia using chemometric methods