Detection of Industrial Process Anomalies Using Isolation Forest and DBSCAN in a Real Manufacturing Environment
Year: 2025
Type: conference-paper
Abstract: Modern manufacturing requires robust methods for monitoring complex industrial systems to ensure operational efficiency and safety. This paper presents a comparative analysis of two distinct unsupervised machine learning algorithms—Isolation Forest and Density-Based Spatial Clustering of Applications with Noise (DBSCAN)—for anomaly detection in industrial environments. We evaluate their performance in detecting faults and defining operational states using real-world industrial datasets. The study leverages a data-driven strategy to address issues such as fault monitoring and process optimization. Results indicate that while Isolation Forest is highly effective for identifying outliers in high-dimensional data (Accuracy 0.91), DBSCAN provides superior interpretability for clustering distinct operational modes and defining signal segmentation. This research contributes to the broader implementation of Digital Twins and Industry 4.0 standards.
Institution: Al-Farabi Kazakh National University
Language: English
Subfield: Control and Systems Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7169843156