Please use this identifier to cite or link to this item: http://dx.doi.org/10.25673/125461
Title: Prediction of PM2.5 and PM10 Concentrations Using Machine and Deep Learning Techniques : Evidence from Tashkent, Uzbekistan
Author(s): Akhatkulov, Sokhobiddin
Yalgoshev, Islom
Uddin, Izhar
Abdullayev, Abubakir
Sariyev, Shohkrukh
Dusanov, Rakhim
Granting Institution: Hochschule Anhalt
Issue Date: 2026
Extent: 1 Online-Ressource (8 Seiten)
Language: English
Abstract: This study investigates several machine and deep learning models to predict the concentrations of PM2.5 and PM10 particles in the air of Tashkent city. These particulate matters are the most influential factors in Tashkent’s high air quality index. The dataset utilized in this investigation comprises PM2.5 and PM10 concentration levels alongside meteorological indicators, including temperature, humidity , wind speed, and air pressure, gathered from 16 monitoring stations around Tashkent city and accessible at https://opendata.tashkent.uz/. The algorithms, Linear Regression (LR), Random Forest (RF), Support Vector Regression (SVR) and Gated Recurrent Unit (GRU) neural network were used to build predictive models on the concentrations of particles PM2.5 and PM10. The model performances were ev aluated using metrics RMSE, MAE, MAPE, R2 and compared. The results show that the GRU and RF models performed significantly better than LR and SVR with R² values of 0.98 and 0.97, respectively, compared to th e value of 0.91 and 0.93 respectively, achieved by LR and SVR. The GRU model was evaluated as the most effective approach due to its ability to deeply explore dynamic relationships across time series.
URI: https://opendata.uni-halle.de//handle/1981185920/127395
http://dx.doi.org/10.25673/125461
Open Access: Open access publication
License: (CC BY-SA 4.0) Creative Commons Attribution ShareAlike 4.0(CC BY-SA 4.0) Creative Commons Attribution ShareAlike 4.0
Appears in Collections:International Conference on Applied Innovations in IT (ICAIIT)

Files in This Item:
File Description SizeFormat 
ICAIIT_2026_14(4)-7-20.pdf1.33 MBAdobe PDFThumbnail
View/Open