Risang Bayu Firdaush
Sekolah Tinggi Meteorologi Klimatologi dan Geofisika

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

PREDIKSI INDEKS NITROGEN DIOKISDA (NO2) MENGGUNAKAN MODEL NEURALPROPHET STUDI KASUS DKI JAKARTA Jhon Paul Estomihi Togatorop; Risang Bayu Firdaush; Yosafat Donni Haryanto
Indonesian Physics Communication Vol 20, No 3 (2023)
Publisher : Universitas Riau

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31258/jkfi.20.3.215-222

Abstract

Nitrogen dioxide gas (NO2) is one of the air quality parameters that can delay nerve recovery after a stroke. DKI Jakarta as the Capital City of the State of Indonesia continues to experience an increase in population which is marked by an increase in the number of motorized vehicles and infrastructure development. Air quality prediction, especially NO2, is important as an anticipatory step in detecting air pollution, especially if the measuring instrument is damaged. This study uses standard air pollutant index (ISPU) data as a time series from 2018 – 2021 to predict the NO2 index in 2022 using the NeuralProphet model. The NeuralProphet model which was designed with parameters of 1000 epochs, learning rate of 0.10, proportion validation of 0.10, and daily frequency produced MAE and RMSE models from the training data of 5.426610 and 7.977689. MAE validation and validation RMSE from the proportion of 0.10 tasting data were 27.762064 and 35.434227. The prediction of the NO2 index for 365 days resulting from the NeuralProphet model shows that the NO2 index experiences an increasing trend which is influenced by seasonal events both annually and weekly. Affecting annual seasons, such as national holidays and monsoon rain patterns. The national holidays in question, such as New Year’s Day, Lunar New Year’s Day, and Christmas Day, trigger an increase in traffic flow. The peak of the NO2 index occurred in February and December, while the NO2 index weakened as it entering October.