Predicting particulate matter PM2.5 using the exponential smoothing and Seasonal ARIMA with R studio

Abstract

In general, public awareness of air quality in Indonesia is increasing. In accordance with the average concentration of particular PM2.5, air quality in Indonesia has improved from 2020 to 2021. However, in some densely populated cities, poor air quality still occurs continuously, for example Jakarta. PM2.5 pollution prediction will be made using monthly data with a case study Jakarta using the time series method, Exponential Smoothing and Seasonal ARIMA model in R studio. In accordance with the analysis, it is found that the Triple Exponential Smoothing and Seasonal ARIMA (0,1,1)(1,1,1)(12) were chosen to be the selected models. When compared to the actual data, they fluctuate move following the actual data. However, when viewed from the average percentage difference with the actual data, the model whose data is close to the actual data is the Seasonal ARIMA model. It is hoped that by knowing the PM2.5 particulate time series model in Jakarta, it can be used as consideration for predicting the monthly average PM2.5 so that it can be seen which months have the highest PM2.5. So that some people who experience respiratory problems, can predict the highest PM2.5 condition and can anticipate early in case of unhealthy air conditions.

The Effect of Junior High School Vocabulary Mastery Using Word Wall Web

“…According to Jannah & Syafryadin, (2022) Word wall serves as an ideal complementary tool for educating young learners, overcoming their tendency to be bored and limited attention span during learning. Amelia et al, (2022) Research supports the idea that Word Wall online platform has a great impact on students' vocabulary acquisition.…”

Prediction of Particulate Matter PM2.5 Using Bidirectional Gated Recurrent Unit with Feature Selection

“…The result shows that MLR achieved 92% and ARIMA achieved 95% accurate prediction of air quality. Amelia et al (2022) developed a PM2.5 prediction model using seasonal ARIMA and triple exponential smoothing time series models. The experimental result shows the seasonal ARIMA produced better prediction than exponential smoothing.…”