Please use this identifier to cite or link to this item:
http://repository.ipb.ac.id/handle/123456789/168611| Title: | Prediksi Kecepatan Getaran Mesin Menggunakan Algoritma Light Gradient Boosting Machine di PT Amerta Indah Otsuka |
| Other Titles: | Machine Vibration Speed Prediction using the Light Gradient Boosting Machine Algorithm at PT Amerta Indah Otsuka |
| Authors: | Renanti, Medhanita Dewi Fawwaz, Kemas muhammad Adnan fakhri Sjaf |
| Issue Date: | 2025 |
| Publisher: | IPB University |
| Abstract: | PT Amerta Indah Otsuka masih melakukan maintenance secara preventif,
yang beresiko menyebabkan downtime yang tidak terduga dan kerugian yang besar.
Penelitian ini bertujuan untuk mengimplementasikan algoritma LightGBM untuk
prediksi kecepatan getaran mesin sebagai fitur predictive maintenance
menggunakan model pembelajaran mesin. Metode CRISP-DM diterapkan karena
memungkinkan pengembangan juga fokus terhadap datasetnya. Hasil penelitian ini
mencakup pemahaman data, pembersihan data, rekayasa fitur, pelatihan model,
evaluasi dan perbandingan model. Algoritma light gradient boosting machine
digunakan sebagai model pembelajaran mesin. Evaluasi model dilakukan
menggunakan mean actual error (MAE). Pembuatan model dilakukan
menggunakan library LigthtGBM dan dashboard dibuat menggunakan FastAPI,
Typescript, dan React. Model sudah menghasilkan akurasi yang cukup baik dan
sesuai berdasarkan hasil metrik yang didapatkan. Diharapkan dengan fitur ini
downtime dapat dihindari, kerugian perusahaan juga dapat diminimalisir. PT Amerta Indah Otsuka still performs preventive maintenance, which risks unexpected downtime and significant losses. This study aims to implement the LightGBM algorithm for machine vibration velocity prediction as a predictive maintenance feature using a machine learning model. The CRISP-DM method was applied because it allows for focused development of the dataset. The research results include data understanding, data cleaning, feature engineering, model training, evaluation, and model comparison. The Light Gradient Boosting Machine algorithm was used as the machine learning model. Model evaluation was performed using the mean actual error (MAE). The model was built using the LightGBM library, and the dashboard was created using FastAPI, TypeScript, and React. The model has produced fairly good accuracy and is consistent based on the obtained metrics. It is hoped that this feature can prevent downtime and minimize company losses. |
| URI: | http://repository.ipb.ac.id/handle/123456789/168611 |
| Appears in Collections: | UT - Software Engineering Technology |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| cover_J0303211171_64a32cd398074c09a131e4e3cf47af5b.pdf | Cover | 2.25 MB | Adobe PDF | View/Open |
| fulltext_J0303211171_a63e5829ae7244efa14c561f43489da5.pdf Restricted Access | Fulltext | 3.04 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.