Skip to content
SDGs Center UDINUS

Evidence record · UDINUS · SDGs

Random Forest Enhancement using Sequential Feature Selection for Machine Maintenance Prediction

Rektorat UDINUS 2025 Published 2026-10-30

Evidence record · UDINUS · SDGs

Unit
Rektorat UDINUS
Period
2025
Goals
Record
585caae6-1ca1-4675-93f2-b73d8f0cd82e
Published
2026-10-30

Universitas Dian Nuswantoro conducted a research activity on the use of the Random Forest algorithm for binary machine failure classification. The activity involved Muljono, Muljono; Arisani, Muhammad Irfan; and Hakim, Ermanu Azizul as participants. The study addressed the use of advanced machine learning, including ensemble learning, in the manufacturing industry. Its focus was to enhance model performance through sequential feature selection, or SFS, for binary machine failure classification.

The activity used a machine dataset with extreme imbalance characteristics. Feature engineering was performed on several independent variables. Pure and formed variables were then selected and separated using a stratified ratio of 65:35 for training and testing data. The processed dataset was trained using the Random Forest algorithm. The study compared Random Forest models combined with sequential feature selection using the backward elimination method and models without sequential feature selection.

Model performance was evaluated using accuracy, recall, precision, F1-score, and Matthews correlation coefficient metrics. The study also presented ROC-AUC and precision-recall curves. The experimental results showed a 15.97 percent increase in recall, from 62.18 to 78.15 percent, compared with the pure Random Forest algorithm. The results indicate potential for further optimization in real-world applications. The activity provides evidence of research involving machine learning, feature selection, machine failure classification, and model performance evaluation in the manufacturing industry.

Reference https://www.scopus.com/pages/publications/105031590339?origin=resultslist