<![CDATA[Industry 4.0 revolutionizes modern manufacturing by enabling the active integration of smart sensors and machine learning to optimize product quality control systems. This research focuses on classifying product quality in the Laser Metal Deposition (LMD) process by applying three machine learning algorithms, namely Decision Tree (DT), Random Forest (RF), and Support Vector Machine (SVM). The dataset consists of four numerical sensor variables, including Optical Sensor, Laser Power, Pressure, and Temperature, with Defect Label as the binary target variable. The Synthetic Minority Oversampling Technique (SMOTE) is used to balance the class distribution. Correlation analysis reveals weak linear relationships among all variables, suggesting the presence of complex non-linear interactions. The Random Forest model produces the best performance with accuracy of 0.88, recall of 0.79, and AUC of 0.80, outperforming Decision Tree and SVM. These findings indicate that ensemble-based methods effectively capture complex patterns within sensor data and offer reliable predictions for quality control in metal manufacturing industries, particularly within Laser Metal Deposition processes.]]>
Copyrights © 2025
