PURMALA, YULIO AGEFA (2022) ANALISIS PREDIKSI WAKTU PERBAIKAN MESIN MENGGUNAKAN MACHINE LEARNING PADA MANUFAKTUR ALAS KAKI DI INDONESIA. S2 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
01 Cover.pdf Download (814kB) | Preview |
|
|
Text (ABSTRAK)
02 Abstrak.pdf Download (666kB) | Preview |
|
![]() |
Text (BAB I)
03 Bab 1.pdf Restricted to Registered users only Download (818kB) |
|
![]() |
Text (BAB II)
04 Bab 2.pdf Restricted to Registered users only Download (1MB) |
|
![]() |
Text (BAB III)
05 Bab 3.pdf Restricted to Registered users only Download (847kB) |
|
![]() |
Text (BAB IV)
06 Bab 4.pdf Restricted to Registered users only Download (1MB) |
|
![]() |
Text (BAB V)
07 Bab 5.pdf Restricted to Registered users only Download (818kB) |
|
![]() |
Text (BAB VI)
08 Bab 6.pdf Restricted to Registered users only Download (637kB) |
|
![]() |
Text (DAFTAR PUSTAKA)
09 Daftar Pustaka.pdf Restricted to Registered users only Download (650kB) |
|
![]() |
Text (LAMPIRAN)
10 Lampiran.pdf Restricted to Registered users only Download (2MB) |
Abstract
Machine breakdowns in the production line are remarkably high because machine repair that needs a lot of time conducted on the production line, not in the machine warehouse. In addition to unbalanced technical skills, this is also because technicians still tend to rely on experience and intuition in estimating repair time intervals. Historical machine breakdown data is digitally recorded through the Andon system, but it is still not being used adequately to aid decision-making. This study introduces analysis of historical machine breakdowns data to provide predictions of repair time intervals with a focus on finding the best algorithm accuracy using a machine learning approach. Using five algorithms of supervised machine learning classification methods: logistic regression, naive Bayes, random forest, k-nearest neighbor, and support vector machine, then evaluate the performance of each algorithm to get the best model. The results of this study prove that historical data on machine breakdown can be used to predict the time interval for machine repair. This study shows that with a standard repair time interval of 18 minutes, the accuracy of the Logistic Regression (LR) algorithm is slightly better than other algorithms. Based on the performance evaluation metric of Receiver Operating Characteristic – Area Under Curve (ROC-AUC), the quality value of the accuracy of the LR model is satisfactory with a percentage of 69% and a difference of 0.5% between the training and test data. Keywords: Industry 4.0, Machine Learning, Classification, Evaluation Metrics, Machine Breakdown Kerusakan mesin yang terjadi di lini produksi sangat tinggi, karena seharusnya perbaikan yang membutuhkan banyak waktu dilakukan di gudang mesin. Selain belum meratanya keahlian teknisi, hal ini juga dikarenakan masih cenderung mengandalkan pengalaman dan intuisi dalam memperkirakan interval waktu perbaikan. Data historis kerusakan mesin tercatat secara digital melalui sistem Andon, namun data tersebut masih belum dapat dimanfaatkan secara baik untuk membantu pengambilan keputusan. Penelitian ini bertujuan untuk melakukan analisis data historis kerusakan mesin untuk memberikan prediksi interval waktu perbaikan kerusakan mesin dengan fokus mencari akurasi algoritma terbaik menggunakan pendekatan machine learning. Menggunakan lima algoritma metode klasifikasi supervised machine learning: logistic regression, naïve bayes, random forest, k-nearest neighbor dan support vector machine, kemudian dilakukan evaluasi performa dari setiap algoritma untuk mendapatkan model terbaik. Hasil penelitian ini membuktikan bahwa data historis kerusakan mesin dapat digunakan untuk melakukan prediksi interval waktu perbaikan kerusakan mesin. Penelitian ini menunjukkan bahwa dengan standar interval waktu perbaikan 18 menit, hasil akurasi algoritma Logistic Regression (LR) sedikit lebih baik dari pada algoritma lainnya. Berdasarkan metrik evaluasi kinerja Receiver Operating Characteristic – Area Under Curve (ROC-AUC), nilai kualitas akurasi model LR adalah memuaskan dengan persentase 69% dan selisih 0,5% antara data latih dan uji. Kata Kunci: Industri 4.0, Machine Learning, Klasifikasi, Metrik Evaluasi, Kerusakan Mesin.
Actions (login required)
![]() |
View Item |