Menghilangkan Autokorelasi Pada Diagram Kontrol Shewhart Menggunakan Diagram Kontrol Residual Berdasarkan Model Extention Support Vector Regression

Bisri, Hasan (2018) Menghilangkan Autokorelasi Pada Diagram Kontrol Shewhart Menggunakan Diagram Kontrol Residual Berdasarkan Model Extention Support Vector Regression. Masters thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
02411650010002-Master_Thesis.pdf - Published Version
Restricted to Repository staff only

Download (4MB) | Request a copy

Abstract

Kualitas merupakan faktor kunci yang mengarahkan kepada keberhasilan, pertumbuhan, dan daya saing bisnis. Kualitas juga merupakan salah satu faktor penting dalam pengambilan keputusan konsumen dalam pemilihan produk dan layanan. Guna meningkatkan kualitas produk dapat memanfaatkan beberapa cara, salah satunya adalah menerapkan statistical process control (SPC). Salah satu tool SPC yang paling banyak diterapkan adalah diagram kontrol yang berguna untuk mengetahui variansi dari proses. Diagram kontrol didasarkan pada asumsi bahwa data mengikuti distribusi normal dan tidak terdapat hubungan antara pengamatan yang berurutan (autokorelasi). Namun dalam proses industri kontinyu kebanyakan data bersifat autokorelasi. Agar bisa menggunakan diagram kontrol secara efektif, autokorelasi dalam data harus dihilangkan. Langkah yang dapat dilakukan untuk pengendalian kualitas pada data autokorelasi adalah dengan memetakan residual hasil pemodelan menggunakan metode time series pada diagram kontrol. Pada penelitian ini dikembangkan diagram kontrol residual berdasarkan model extention Support vector regression yaitu Least square support vector regression dan Genetic algorithm support vector regression untuk mengatasi kasus autokorelasi pada proses. Kriteria kebaikan model dalam penelitian ini menggunakan nilai Root Mean Square Error (RMSE). Semakin kecil nilai RMSE maka model yang digunakan semakin baik. Setelah dilakukan perhitungan menggunakan metode regresi, Support vector regression dan metode Extention support vector regression, metode yang paling baik adalah Genetic algorithm support vector regression berdasarkan nilai RMSE sebesar 1,554310 dan 0,5565. ============ Quality is a key factor that leads to business success, growth, and competitiveness. Quality is also an important factor in consumer decision making in the selection of products and services. In order to improve product quality can utilize several ways, one of them is apply statistical process control (SPC). One of the most widely applied SPC tools is the control chart which is useful for knowing the variance of the process. The control chart is based on the assumption that data follows a normal distribution and there is no relationship between successive observations (autocorrelation). But in the process of continuous industry most data are autocorrelation. In order to use the control chart effectively, autocorrelation in the data must be eliminated. Steps that can be done to control the quality of the autocorrelation data is to map the residual results of modeling using time series method in the control chart. In this research, the residual control charts are developed based on the extension support vector regression model that is Least square support vector regression and Genetic algorithm support vector regression to overcome the case of autocorrelation in the process. Criteria of model goodness in this research use Root Mean Square Error (RMSE). The smaller the value of RMSE then the model used the better. After calculation using regression method, Support vector regression and Extension support vector regression method, the best method is Genetic algorithm support vector regression based on RMSE value of 1.554310 and 0.5565.

Item Type: Thesis (Masters)
Uncontrolled Keywords: Autokorelasi; Diagram Kontrol; Least Square; Genetic Algorithm; Support Vector Regression; Diagram Kontrol Residual
Subjects: Q Science > QA Mathematics > QA278.2 Regression Analysis
T Technology > T Technology (General) > T57.5 Data Processing
T Technology > TS Manufactures > TS156 Quality Control. QFD
Divisions: Faculty of Industrial Technology > Industrial Engineering > (S2) Master Theses
Depositing User: Hasan Bisri
Date Deposited: 12 Jul 2018 07:33
Last Modified: 12 Jul 2018 07:33
URI: http://repository.its.ac.id/id/eprint/52203

Actions (login required)

View Item View Item