Bayu, M Septian Krisna Bayu (2026) Pengendalian Kualitas Udara Menggunakan Exponentially Weighted Moving Average (EWMA) Berbasis Residual Regresi Dinamis Terhadap Kondisi Cuaca Di Kota Surabaya. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5002211062-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5002211062-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (4MB) | Request a copy |
Abstract
Tugas akhir ini membahas tentang pengendalian kualitas udara menggunakan Exponentially Weighted Moving Average (EWMA) berbasis residual regresi dinamis terhadap kondisi cuaca di Kota Surabaya. Kualitas udara pada Tugas Akhir ini direpresentasikan melalui konsentrasi polutan PM2.5, sedangkan kondisi cuaca direpresentasikan oleh suhu udara dan kecepatan angin. Tujuan Tugas Akhir ini adalah menganalisis hubungan dinamis antara faktor cuaca terhadap konsentrasi PM2.5 dengan menerapkan sistem pemantauan Control Chart Exponentially Weighted Moving Average (EWMA) berbasis residual untuk menganalisis Out of Control (OOC). Hubungan antara konsentrasi PM2.5 terhadap suhu udara dan kecepatan angin dimodelkan menggunakan Regresi Dinamis yaitu Autoregressive Distributed Lag (ARDL). Hasil Tugas Akhir ini menunjukkan bahwa Model ARDL [1], [6], [12]; [1]; [1] merupakan model terbaik dengan parameter yang diestimasi menggunakan Weighted Least Squared (WLS). Bobot (w) yang digunakan pada Weighted Least Aquared (WLS) adalah w = 1/|ˆε| dimana nilai ˆε diperoleh dari hasil estimasi parameter model menggunakan Ordinary Least Squared (OLS). Model ARDL [1], [6], [12]; [1]; [1] menunjukkan kinerja yang sangat kuat dengan nilai R2 sebesar 0,99 dan tingkat akurasi yang tinggi dengan MAPE sebesar 5,39%. Residual Model ARDL [1], [6], [12]; [1]; [1] digunakan sebagai input pengendalian kualitas udara menggunakan EWMA. Metode EWMA diterapkan dengan nilai pembobot (λ) mulai dari 0,1 hingga 0,4. Rentang ini dipilih untuk melihat kemampuan deteksi yang berbeda yaitu nilai 0,1 untuk menangkap peningkatan polusi yang bergerak pelan, sedangkan 0,4 digunakan untuk merespons lonjakan polusi yang lebih cepat. Hasil menunjukkan bahwa EWMA mendeteksi adanya Out of Control (OOC) pada data testing untuk λ = 0, 3 dan 0, 4. Perbandingan EWMA pada dataset keseluruhan menunjukkan jumlah OOC pada EWMA berbasis residual selalu lebih rendah daripada data mentah. Hal ini menegaskan bahwa EWMA berbasis residual ARDL efektif dalam mengeliminasi variasi yang disebabkan oleh faktor cuaca yang telah termodelkan, sehingga OOC yang tersisa merepresentasikan pengaruh faktor peningkatan kandungan PM2.5 selain suhu udara dan kecepatan angin.
================================================================================================================================
This study investigates air quality control in Surabaya City using an Exponentially Weighted Moving Average (EWMA) control chart based on dynamic regression residuals accounting for weather conditions. Air quality is represented by PM2.5 pollutant concentration, while weather conditions are represented by air temperature and wind speed. The objective of this study is to analyze the dynamic relationship between weather factors and PM2.5 concentration and to apply a residual-based Exponentially Weighted Moving Average (EWMA) Control Chart monitoring system to detect Out of Control (OOC) signals. The relationship between PM2.5 concentration, air temperature, and wind speed is modeled using the Dynamic Regression method, specifically the Autoregressive Distributed Lag (ARDL). The results indicate that the ARDL [1], [6], [12]; [1]; [1] model is the best-fitting model, with parameters estimated using Weighted Least Squares (WLS). The weight (w) used in Weighted Least Squares (WLS) is defined as w = 1/|ˆε| , where ˆε is obtained from the initial parameter estimation using Ordinary Least Squares (OLS). The ARDL [1], [6], [12]; [1]; [1] model exhibits robust performance with an R2 value of = 0.99 and a high degree of accuracy, indicated by a MAPE of 5.39%. The residuals from the ARDL [1], [6], [12]; [1]; [1] model serve as the input for air quality control using EWMA. The EWMA method is applied using smoothing parameters (λ) ranging from 0.1 to 0.4. This range was selected to evaluate varying detection capabilities: a value of 0.1 is used to capture gradual pollution drifts, while 0.4 is used to respond to more rapid pollution spikes. The results demonstrate that EWMA detects Out of Control (OOC) signals in the testing data for λ = 0.3 and 0.4. A comparison of EWMA performance across the entire dataset reveals that the number of OOC signals in the residual-based EWMA is consistently lower than that of the raw data. This confirms that the ARDL residual-based EWMA is effective in eliminating variations attributed to modeled weather factors; thus, the remaining OOC signals represent influences on PM2.5 levels other than air temperature and wind speed.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | PM2.5, Cuaca, Regresi Dinamis, EWMA, PM2.5, Wheater, Dynamic Regression, EWMA |
| Subjects: | Q Science > QA Mathematics > QA275 Theory of errors. Least squares. Including statistical inference. Error analysis (Mathematics) Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis |
| Depositing User: | M. Septian Krisna Bayu |
| Date Deposited: | 23 Jan 2026 01:14 |
| Last Modified: | 23 Jan 2026 01:14 |
| URI: | http://repository.its.ac.id/id/eprint/130140 |
Actions (login required)
 |
View Item |