Potential of ARIMA-ANN, ARIMA-SVM, DT and CatBoost for Atmospheric PM<sub>2.5</sub> Forecasting in Bangladesh

Atmospheric particulate matter (PM) has major threats to global health, especially in urban regions around the world. Dhaka, Narayanganj and Gazipur of Bangladesh are positioned as top ranking polluted metropolitan cities in the world. This study assessed the performance of the application of hybrid...

Full description

Bibliographic Details
Main Authors: Shihab Ahmad Shahriar, Imrul Kayes, Kamrul Hasan, Mahadi Hasan, Rashik Islam, Norrimi Rosaida Awang, Zulhazman Hamzah, Aweng Eh Rak, Mohammed Abdus Salam
Format: Article
Language:English
Published: MDPI AG 2021-01-01
Series:Atmosphere
Subjects:
Online Access:https://www.mdpi.com/2073-4433/12/1/100
id doaj-d41c94813f6e433bbe002aaf53dab756
record_format Article
spelling doaj-d41c94813f6e433bbe002aaf53dab7562021-01-13T00:02:04ZengMDPI AGAtmosphere2073-44332021-01-011210010010.3390/atmos12010100Potential of ARIMA-ANN, ARIMA-SVM, DT and CatBoost for Atmospheric PM<sub>2.5</sub> Forecasting in BangladeshShihab Ahmad Shahriar0Imrul Kayes1Kamrul Hasan2Mahadi Hasan3Rashik Islam4Norrimi Rosaida Awang5Zulhazman Hamzah6Aweng Eh Rak7Mohammed Abdus Salam8Department of Environmental Science and Disaster Management, Noakhali Science and Technology University, Noakhali 3814, BangladeshDepartment of Environmental Science and Disaster Management, Noakhali Science and Technology University, Noakhali 3814, BangladeshDepartment of Environmental Science and Disaster Management, Noakhali Science and Technology University, Noakhali 3814, BangladeshDepartment of Environmental Science and Disaster Management, Noakhali Science and Technology University, Noakhali 3814, BangladeshDepartment of Computer Science and Media, Beuth University of Applied Sciences, 13353 Berlin, GermanyFaculty of Earth Science, Jeli Campus, Universiti Malaysia Kelantan, Jeli 17600, Kelantan, MalaysiaFaculty of Earth Science, Jeli Campus, Universiti Malaysia Kelantan, Jeli 17600, Kelantan, MalaysiaFaculty of Earth Science, Jeli Campus, Universiti Malaysia Kelantan, Jeli 17600, Kelantan, MalaysiaDepartment of Environmental Science and Disaster Management, Noakhali Science and Technology University, Noakhali 3814, BangladeshAtmospheric particulate matter (PM) has major threats to global health, especially in urban regions around the world. Dhaka, Narayanganj and Gazipur of Bangladesh are positioned as top ranking polluted metropolitan cities in the world. This study assessed the performance of the application of hybrid models, that is, Autoregressive Integrated Moving Average (ARIMA)-Artificial Neural Network (ANN), ARIMA-Support Vector Machine (SVM) and Principle Component Regression (PCR) along with Decision Tree (DT) and CatBoost deep learning model to predict the ambient PM<sub>2.5</sub> concentrations. The data from January 2013 to May 2019 with 2342 observations were utilized in this study. Eighty percent of the data was used as training and the rest of the dataset was employed as testing. The performance of the models was evaluated by R<sup>2</sup>, RMSE and MAE value. Among the models, CatBoost performed best for predicting PM<sub>2.5</sub> for all the stations. The RMSE values during the test period were 12.39 µg m<sup>−3</sup>, 13.06 µg m<sup>−3</sup> and 12.97 µg m<sup>−3</sup> for Dhaka, Narayanganj and Gazipur, respectively. Nonetheless, the ARIMA-ANN and DT methods also provided acceptable results. The study suggests adopting deep learning models for predicting atmospheric PM<sub>2.5</sub> in Bangladesh.https://www.mdpi.com/2073-4433/12/1/100air pollutionPM<sub>2.5</sub>ARIMA-ANNARIMA-SVMCatBoostdeep learning model
collection DOAJ
language English
format Article
sources DOAJ
author Shihab Ahmad Shahriar
Imrul Kayes
Kamrul Hasan
Mahadi Hasan
Rashik Islam
Norrimi Rosaida Awang
Zulhazman Hamzah
Aweng Eh Rak
Mohammed Abdus Salam
spellingShingle Shihab Ahmad Shahriar
Imrul Kayes
Kamrul Hasan
Mahadi Hasan
Rashik Islam
Norrimi Rosaida Awang
Zulhazman Hamzah
Aweng Eh Rak
Mohammed Abdus Salam
Potential of ARIMA-ANN, ARIMA-SVM, DT and CatBoost for Atmospheric PM<sub>2.5</sub> Forecasting in Bangladesh
Atmosphere
air pollution
PM<sub>2.5</sub>
ARIMA-ANN
ARIMA-SVM
CatBoost
deep learning model
author_facet Shihab Ahmad Shahriar
Imrul Kayes
Kamrul Hasan
Mahadi Hasan
Rashik Islam
Norrimi Rosaida Awang
Zulhazman Hamzah
Aweng Eh Rak
Mohammed Abdus Salam
author_sort Shihab Ahmad Shahriar
title Potential of ARIMA-ANN, ARIMA-SVM, DT and CatBoost for Atmospheric PM<sub>2.5</sub> Forecasting in Bangladesh
title_short Potential of ARIMA-ANN, ARIMA-SVM, DT and CatBoost for Atmospheric PM<sub>2.5</sub> Forecasting in Bangladesh
title_full Potential of ARIMA-ANN, ARIMA-SVM, DT and CatBoost for Atmospheric PM<sub>2.5</sub> Forecasting in Bangladesh
title_fullStr Potential of ARIMA-ANN, ARIMA-SVM, DT and CatBoost for Atmospheric PM<sub>2.5</sub> Forecasting in Bangladesh
title_full_unstemmed Potential of ARIMA-ANN, ARIMA-SVM, DT and CatBoost for Atmospheric PM<sub>2.5</sub> Forecasting in Bangladesh
title_sort potential of arima-ann, arima-svm, dt and catboost for atmospheric pm<sub>2.5</sub> forecasting in bangladesh
publisher MDPI AG
series Atmosphere
issn 2073-4433
publishDate 2021-01-01
description Atmospheric particulate matter (PM) has major threats to global health, especially in urban regions around the world. Dhaka, Narayanganj and Gazipur of Bangladesh are positioned as top ranking polluted metropolitan cities in the world. This study assessed the performance of the application of hybrid models, that is, Autoregressive Integrated Moving Average (ARIMA)-Artificial Neural Network (ANN), ARIMA-Support Vector Machine (SVM) and Principle Component Regression (PCR) along with Decision Tree (DT) and CatBoost deep learning model to predict the ambient PM<sub>2.5</sub> concentrations. The data from January 2013 to May 2019 with 2342 observations were utilized in this study. Eighty percent of the data was used as training and the rest of the dataset was employed as testing. The performance of the models was evaluated by R<sup>2</sup>, RMSE and MAE value. Among the models, CatBoost performed best for predicting PM<sub>2.5</sub> for all the stations. The RMSE values during the test period were 12.39 µg m<sup>−3</sup>, 13.06 µg m<sup>−3</sup> and 12.97 µg m<sup>−3</sup> for Dhaka, Narayanganj and Gazipur, respectively. Nonetheless, the ARIMA-ANN and DT methods also provided acceptable results. The study suggests adopting deep learning models for predicting atmospheric PM<sub>2.5</sub> in Bangladesh.
topic air pollution
PM<sub>2.5</sub>
ARIMA-ANN
ARIMA-SVM
CatBoost
deep learning model
url https://www.mdpi.com/2073-4433/12/1/100
work_keys_str_mv AT shihabahmadshahriar potentialofarimaannarimasvmdtandcatboostforatmosphericpmsub25subforecastinginbangladesh
AT imrulkayes potentialofarimaannarimasvmdtandcatboostforatmosphericpmsub25subforecastinginbangladesh
AT kamrulhasan potentialofarimaannarimasvmdtandcatboostforatmosphericpmsub25subforecastinginbangladesh
AT mahadihasan potentialofarimaannarimasvmdtandcatboostforatmosphericpmsub25subforecastinginbangladesh
AT rashikislam potentialofarimaannarimasvmdtandcatboostforatmosphericpmsub25subforecastinginbangladesh
AT norrimirosaidaawang potentialofarimaannarimasvmdtandcatboostforatmosphericpmsub25subforecastinginbangladesh
AT zulhazmanhamzah potentialofarimaannarimasvmdtandcatboostforatmosphericpmsub25subforecastinginbangladesh
AT awengehrak potentialofarimaannarimasvmdtandcatboostforatmosphericpmsub25subforecastinginbangladesh
AT mohammedabdussalam potentialofarimaannarimasvmdtandcatboostforatmosphericpmsub25subforecastinginbangladesh
_version_ 1724339808138952704