Image of METODE ENSEMBLE MACHINE LEARNING PREDIKSI KEKASARAN PERMUKAAN BAJA S45C PADA PROSES CNC MILLING
Bookmark Share

Skripsi

METODE ENSEMBLE MACHINE LEARNING PREDIKSI KEKASARAN PERMUKAAN BAJA S45C PADA PROSES CNC MILLING

Surface roughness is a critical quality indicator in precision manufacturing industries due to its direct impact on fatigue resistance and tribological performance of machine components. This study aims to predict surface roughness (Ra) values in the CNC milling process of S45C steel using flood cooling by integrating Support Vector Regression (SVR) and Gaussian Process Regression (GPR) algorithms. The integration is achieved through Ensemble Learning approaches, specifically Weighted Average and Stacking, with hyperparameters automatically optimized using the Optuna framework. Based on 119 experimental data points, feature importance analysis identified feed rate as the most influential variable, contributing 68.5 percent. Model evaluation results showed that the SVR RBF kernel outperformed GPR individually (MAE 0.2035, MSE 0.0749, RMSE 0.2737, and R^2 0.5294), while GPR recorded lower performance (MAE 0.2233, MSE 0.0850, RMSE 0.2916, and R^2 0.4658). However, the application of ensemble techniques drastically improved prediction accuracy. The Weighted Ensemble yielded (MAE 0.1491, MSE 0.0435, RMSE 0.2087, and R^2 0.9419), while the best performance was achieved by the Stacked Ensemble (MAE 0.1501, MSE 0.0381, RMSE 0.1952, and R^2 0.9491). These findings validate that ensemble techniques provide significantly higher stability and precision compared to single models to support production efficiency.

Availability
#
Central Library (Reference) T1918192025
T191819
Available but not for loan - Not for Loan
Detail Information
Series Title
-
Call Number
T1918192025
Publisher
Indrlaya : Prodi Teknik Mesin, Fakultas Teknik Universitas Sriwijaya.,
Collation
xxx, 65 hlm.; ilus.; 29 cm
Language
Indonesia
ISBN/ISSN
-
Classification
621.910 7
Content Type
Text
Media Type
unmediated
Carrier Type
-
Edition
-
Subject(s)
Prodi Teknik Mesin
Specific Detail Info
-
Statement of Responsibility
Other version/related
File Attachment
Comments
You must be logged in to post a comment