sulistiawanti, nur (2022) Penerapan Diagram Kontrol Multivariate Exponentially Weighted Moving Average (MEWMA) Dan Multivariate Exponentially Weighted Moving Variance (MEWMV) Berbasis Residual XGBoost Regression Pada Pengendalian Kualitas Air Produksi PAM Tirta Mangkaluku Di Kota Palopo. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 06211840000068-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
06211840000068-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (3MB) | Request a copy |
Abstract
PAM Tirta Mangkaluku Kota Palopo merupakan salah satu Badan Usaha Milik Daerah (BUMD) yang dimiliki oleh pemerintah Kota Palopo yang bergerak di bidang jasa penyediaan dan pelayanan air bersih bagi masyarakat Kota Palopo. Terdapat 5 Instalasi Penjernihan Air Minum (IPAM) yang dikelola oleh PAM Tirta Mangkaluku Kota Palopo, terdiri dari IPAM I Abdul Majid, IPAM II Magandang, IPAM III Battang, IPAM IV Latuppa, dan IPAM V Batupapan. Pengujian kualitas air produksi di setiap IPAM menggunakan beberapa karakteristik kualitas, diantaranya terdapat kekeruhan, pH, dan sisa chlor, dimana antar karakteristik kualitasnya saling berhubungan. Data yang digunakan dalam penelitian ini adalah data sekunder dari Bagian Pengendalian Proses di Kantor PAM Tirta Mangkaluku Kota Palopo. Data sekunder diperoleh dari hasil pengujian kualitas air produksi di setiap IPAM PAM Tirta Mangkaluku Kota Palopo pada bulan Juli-Desember tahun 2021 periode harian. Dalam kondisi karakteristik yang lebih dari satu, penggunaan peta kendali multivariat lebih sesuai dibandingkan dengan peta kendali univariat. Selain itu, data dengan tingkat autokorelasi yang tinggi dapat menggunakan diagram kontrol berbasis residual untuk mengatasi autokorelasi. XGBoost (Extream Gradient Boosting) merupakan pengembangan dari algoritma Gradient Tree Boosting yang berbasis algoritma ensemble, secara efisien dapat menangani permasalahan machine learning yang berskala besar. Dengan demikian, pada penelitian ini metode yang digunakan untuk mengontrol karakteristik kualitas air produksi PAM Tirta Mangkaluku adalah diagram kontrol multivariat. Diagram kontrol multivariat yang diusulkan adalah diagram kontrol Multivariate Exponentially Weighted Moving Average (MEWMA) untuk mengontrol rata-rata proses dan diagram kontrol Multivariate Exponentially Weighted Moving Variance (MEWMV) untuk mengontrol variabilitas proses berbasis residual dari regresi XGBoost. Pemodelan XGBoost digunakan untuk mengatasi autokorelasi. Nilai residual digunakan dalam pembuatan diagram kontrol MEWMV dan MEWMA. Pada data fase I, proses telah terkendali secara statistik baik variabilitas maupun rata-rata prosesnya dengan pembobot yang digunakan pada analisis variabilitas proses λ = 0,4 dan ω = 0,4, serta pembobotan pada rata-rata proses λ = 0,3. Pada data fase II diperoleh hanya IPAM V Batupapan yang variabilitas prosesnya telah terkendali secara statistik. Untuk IPAM lainnya pada fase II belum terkendali secara statistik baik variabilitas maupun rata-rata prosesnya. Berdasarkan hasil analisis kapabilitas proses hasil pengujian kualitas air produksi pada data fase I, diperoleh nilai Cp lebih dari 1 untuk setiap karakteristik kualitas dan Cpk kurang dari 1 untuk karakteristik kualitas sisa chlor yang menunjukkan kualitas air produksi PAM Tirta Mangkaluku pada prosesnya belum kapabel dengan tingkat akurasi yang kurang baik namun presisinya baik.
==============================================================================================================================
PAM Tirta Mangkaluku Palopo City is one of the Regional Owned Enterprises (BUMD) owned by the Palopo City government which is engaged in providing clean water services for the people of Palopo City. There are 5 Drinking Water Purification Installations (IPAM) managed by PAM Tirta Mangkaluku, Palopo City, consisting of IPAM I Abdul Majid, IPAM II Magandang, IPAM III Battang, IPAM IV Latuppa, and IPAM V Batupapan. Testing the quality of production water at each IPAM uses several quality characteristics, including turbidity, pH, and residual chlorine, where the quality characteristics are interconnected. The data used in this study is secondary data from the Process Control Section at the PAM Tirta Mangkaluku Office, Palopo City. Secondary data was obtained from the results of testing the quality of production water at each IPAM PAM Tirta Mangkaluku Palopo City on July-until December 2021 for a daily period. In conditions of more than one characteristic, the use of a multivariate control chart is more appropriate than a univariate control chart. Besides that, data with level high autocorrelation could use control chart residual based for resolve autocorrelation. XGBoost (Extream Gradient Boosting) is a development from based Gradient Tree Boosting algorithm ensemble algorithm, automatically efficient could handle scalable machine learning problems big. Thus, in this study the method used to control the quality characteristics of the production water at PAM Tirta Mangkaluku is a multivariate control chart. The proposed multivariate control chart is the Multivariate Exponentially Weighted Moving Average (MEWMA) control chart for controlling the process mean and the Multivariate Exponentially Weighted Moving Variance (MEWMV) control chart for controlling process variability based on residual of XGBoost regression. XGBoost modeling is used to overcome autocorrelation. Residual values are used in the manufacture of MEWMV and MEWMA control diagrams. In phase I, the process has been statistically controlled, both the variance and the process mean, with the weights used in the process variability analysis = 0.4 and = 0.4, and the weighting on the process mean is = 0.3. In phase 2 data obtained only IPAM V Batupapan whose process variability has been controlled statistically. For other IPAMs in phase 2, it has not been statistically controlled, both the variability and the process mean. Based on the results of the process capability analysis of the production water quality test results in phase 1 data, obtained a Cp value of more than 1 for each quality characteristic and a Cpk of less than 1 for the quality characteristics of residual chlorine, indicating the quality of the production water at PAM Tirta Mangkaluku in the process is not yet capable with the level of accuracy that is not good but the precision is good.
| Item Type: | Thesis (Other) |
|---|---|
| Additional Information: | RSSt 519.86 Sul p-1 2022 |
| Uncontrolled Keywords: | Autokorelasi, kapabilitas proses, kualitas air, MEWMA, MEWMV, regresi XGBoost. Autocorrelation, process capability, water quality, MEWMA, MEWMV, XGBoost regression. |
| Subjects: | H Social Sciences > HA Statistics |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Mr. Marsudiyana - |
| Date Deposited: | 11 Jun 2026 02:19 |
| Last Modified: | 11 Jun 2026 02:19 |
| URI: | http://repository.its.ac.id/id/eprint/133716 |
Actions (login required)
 |
View Item |