Comparison of Statistical and Deep Learning
Methods for Forecasting PM&lt;sub&gt;2.5&lt;/sub&gt; Concentration
in Northern Thailand

Wongrin, Weerinrada; Chaisee, Kuntalee; Suphawan, Kamonrat

doi:10.15244/pjoes/157072

Cited by 3 publications

(2 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many statistical methods have been used to study and forecast PM concentrations using air pollutants data, such as Holt-Winters exponential smoothing, autoregressive integrated moving average (ARIMA), linear regression model. These methods use collections of probability distribution and assumption to make predictions (Wongrin et al, [10]). In contrast, machine learning is another method to focus on prediction by using learning algorithms to find patterns and can apply to deal with data and make a prediction.…”

Section: Introductionmentioning

confidence: 99%

Application of Deep Belief Network in Weather Modeling: PM2.5 Concentration in Thailand

Wanishsakpong,

Atsawachanakan,

Panityakul

2024

Int. J. Anal. Appl.

View full text Add to dashboard Cite

In Thailand, the number of particles matter with diameter of less than 2.5 microns or PM2.5 concentration exceed the standard in many areas, especially in Chiang Mai. This affects the image of the country in terms of economy, health, and environment. The objective of this research is to study the structure of model for PM2.5 concentration by using a Deep Belief Network (DBN) with the daily data set of PM2.5 concentration from the air quality monitoring station at Yupparaj Wittayalai School, Chiang Mai. The data was analyzed through an unsupervised path using the Minimizing Contrastive Divergence (MCD) algorithm, followed by a supervised path using Back-Propagation Neural Network (BPNN) algorithm to estimate the parameters of DBN. The result shows that the optimal DBN structure has 5 input nodes and 20 hidden neurons in the first hidden layer. This model has an 88.4 percent accuracy in forecasting PM2.5 concentration. In addition, this model can be applied for other weather forecasting such as rainfall or water level in a basin.

show abstract

Section: Introductionmentioning

confidence: 99%

Application of Deep Belief Network in Weather Modeling: PM2.5 Concentration in Thailand

Wanishsakpong,

Atsawachanakan,

Panityakul

2024

Int. J. Anal. Appl.

View full text Add to dashboard Cite

show abstract

“…The Brazilian scholar Ventura et al (2019) [5] proposed using the Holt-Winters model to predict air quality and got better results in the simulation prediction of PM2.5 in the industrial area of Rio de Janeiro. Wongrin et al (2023) [6] used statistical methods and deep learning techniques to predict daily average PM2.5 concentrations in northern Thailand and showed that ARIMA and ETS models performed better than deep learning methods at most stations. Aladağ (2021) [7] combined wavelet transform and ARIMA models to construct a WT-ARIMA model to predict monthly PM10 concentrations in Erzurum, Turkey, and indicated that the hybrid model gave better predictions than the traditional ARIMA model.…”

Section: Introductionmentioning

confidence: 99%

Predicting concentrations of atmospheric particle matters in Guangzhou by time series models

Zhong¹,

Cheng²

2023

Int. J. Stat. Appl. Math.

View full text Add to dashboard Cite

Particulate matter is one of the major air pollutants closely related to human health. In order to predict atmospheric particulate matter concentrations effectively and accurately, this paper utilized ARIMA model, Holt-Winters model, STL-Holt model and STL-ARIMA model to carry out prediction experiments based on hourly PM2.5 and PM10 concentration historical data in Guangzhou city. The results showed that the four models were effective in predicting hourly PM2.5 and PM10 concentrations. The RMSE, MAE, MAPE, and 2 R metrics were used to evaluate the prediction accuracy of the models.It was found that the Holt-Winters model performed best among the four models. This study may provide guides for the environmental authorities in forecasting atmospheric particulate matter concentrations.

show abstract

A Deep Learning Framework for Modeling Temporal Dependencies and Hierarchies in Hourly Electricity Demand Load

Shoko,

Moroke,

Makatjane

2024

Machine Learning and Computer Vision for Renewable Energy

View full text Add to dashboard Cite

The limitations of traditional deep learning models in processing vast volumes of data and modelling complicated temporal dependencies make it difficult to effectively satisfy these objectives for short-term load forecasting (STLF). This chapter utilises deep learning, which enables the following: k-means clustering to comprehend hourly electricity demand load trend, extraction of complex features with non-linear interactions that impact electricity demand load, handling of long-term dependencies through the modelling of temporal hierarchies in the time series data via long short-term memory, and recurrent neural network to capture dependencies across time steps. The performance of the state-of-the-art hourly demand forecast models with the k-means variant is compared with that of the RNN-LSTM-CNN-copula. Furthermore, it is noted that RNN-LSTM-CNN-copula is a viable deep learning model for energy consumption forecast problems because of its capacity to learn the spatio-temporal dependencies in the hourly electrical demand load data, which results in an accurate energy demand forecast.

show abstract

Comparison of Statistical and Deep Learning Methods for Forecasting PM<sub>2.5</sub> Concentration in Northern Thailand

Cited by 3 publications

References 13 publications

Application of Deep Belief Network in Weather Modeling: PM2.5 Concentration in Thailand

Application of Deep Belief Network in Weather Modeling: PM2.5 Concentration in Thailand

Predicting concentrations of atmospheric particle matters in Guangzhou by time series models

A Deep Learning Framework for Modeling Temporal Dependencies and Hierarchies in Hourly Electricity Demand Load

Contact Info

Product

Resources

About