Garuda - Garba Rujukan Digital

Kombinasi Analytical Hierarchy Process, C4.5, dan Particle Swarm Optimization pada Klasifikasi Pegawai Dafiz Adi Nugroho; Catur Edi Widodo; Rahmat Gernowo
JSINBIS (Jurnal Sistem Informasi Bisnis) Vol 12, No 2 (2022): Volume 12 Nomor 2 Tahun 2022
Publisher : Universitas Diponegoro

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21456/vol12iss2pp81-88

Decision Tree C4.5 is widely implemented in various research fields in determining classification, but there are still weaknesses in Decision Tree C4.5, one of which is that it cannot rank each alternative. In this study, to overcome the weakness of Decision Tree C4.5, a combination of Analytical Hierarchy Process (AHP) methods, Decision Tree C4.5, and Particle Swarm Optimization (PSO) methods is proposed in the case study of employee classification for promotion recommendations. The research begins by determining the criteria and weighting criteria from the interview results which are then processed with AHP to produce employee ratings and eligibility labels for the classification process. The classification process uses the Decision Tree C4.5 method which is optimized with the PSO algorithm so as to produce employee eligibility data for promotions. The results of the combined research of AHP, Decision Tree C4.5, and PSO methods show that AHP can produce employee ratings based on performance and potential criteria, and Decision Tree C4.5 classification and optimization with PSO have better accuracy results, namely 95.80% compared to Decision Tree C4.5 method without PSO optimization is 93.40%. Based on the results of the ranking and classification of this research can be used as a basis for promotion of employees.

Hybrid Machine Learning for Early Prediction of At-Risk Students with Imbalanced Data Esti Wijayanti; Widowati Widowati; Catur Edi Widodo
Journal of Applied Data Sciences Vol 7, No 2: May 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i2.1368

The phenomenon of student dropout remains a major challenge for higher education institutions because it impacts academic performance and institutional reputation. Identification of students at risk of dropping out is often hampered by data imbalance, where the number of dropouts is far fewer than active students, so conventional prediction models tend to be biased towards the majority class. This study aims to develop an accurate and reliable prediction framework for students at risk of dropping out to detect at-risk students through a hybrid machine learning approach with data balancing techniques. The main contribution of this study is the integration of Support Vector Machine and Extreme Gradient Boosting in a stacked ensemble architecture supported by data balancing optimization techniques. The proposed model leverages the ability of Support Vector Machine to separate complex classification patterns, while Extreme Gradient Boosting improves prediction accuracy through iterative learning and modeling interactions between variables. The problem of data imbalance is addressed through oversampling techniques for the minority class so that the model learning process becomes more balanced. The model framework is tested using a dataset consisting of 3,652 students with academic, socioeconomic, and behavioral variables. Experimental results show that the proposed hybrid model outperforms the single model, with an accuracy rate of 97 percent, a precision rate of 94 percent, and a recall rate of 95 percent. These findings suggest that a combination of complementary machine learning methods, coupled with data optimization, can significantly improve the predictive ability of student dropout. The practical implication of this research is the availability of a robust decision support system for universities in designing timely and targeted interventions. By identifying students at risk of dropping out, institutions can strengthen retention strategies, improve student academic success, and reduce dropout rates more effectively.

SKIN DISEASE CLASSIFICATION USING EFFICIENT TRANSFER LEARNING AND ATTENTION MECHANISM KURNIA ADI CAHYANTO; KUSWORO ADI; CATUR EDI WIDODO
Rabit : Jurnal Teknologi dan Sistem Informasi Univrab Vol 11 No 1 (2026): Januari
Publisher : LPPM Universitas Abdurrab

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36341/rabit.v11i1.7278

Skin diseases are a common health issue that is often underestimated, as most are mild and can be treated with over-the-counter medications. However, some types, such as melanoma, can be cancerous and deadly if not treated properly. Melanoma is caused by excessive exposure to ultraviolet rays and has a recovery rate of 99% if diagnosed on time, but it decreases to 20% in advanced stages. This study developed a multi-category skin disease classification model using transfer learning through a previously trained model such as EfficientNetV2S with Attention Mechanism to overcome overfitting and improve accuracy. The dataset used is ISIC2019 with 8 classes of skin diseases and 25,331 samples, after data augmentation was performed to increase the sample size. The EffCANet model showed a test accuracy of 94.81%, higher than previous studies, indicating a decrease in the overfitting gap and an improvement in test accuracy results.

Comparative Approaches to Clustering for Profiling Students in Educational Data Mining Noor Azizah; Kusworo Adi; Catur Edi Widodo
Jurnal Informatika dan Rekayasa Perangkat Lunak Vol. 8 No. 1 (2026): Maret
Publisher : Universitas Wahid Hasyim

Show Abstract | Download Original | Original Source | Check in Google Scholar

This study aims to compare the performance of five clustering algorithms, a K-Means, K-Medoids, Fuzzy C-Means (FCM), DBSCAN, and Gaussian Mixture Model (GMM) in profiling 239 students using quantitative data. The methodology includes data collection, refinement, transformation, application of clustering algorithms, and evaluation using the Silhouette Score, Davies–Bouldin Index, and execution time. The results indicate that K-Means provides the most balanced performance, achieving the highest Silhouette score with well-defined cluster separation. K-Medoids and GMM demonstrate competitive performance, while DBSCAN excels in detecting outliers but produces an excessive number of clusters, limiting its interpretability for profiling. FCM performs the weakest due to poor cluster separability. Overall, K-Means is recommended as the primary approach for student profiling, while other algorithms may complement specific analytical needs.

Automated convolutional neural network-based reject–repeat decision system for chest radiographs Dwi Rochmayanti; Kusworo Adi; Catur Edi Widodo
IAES International Journal of Artificial Intelligence (IJ-AI) Vol 15, No 5: October 2026
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijai.v15.i5.pp4121-4132

Image quality in radiography is essential for diagnostic accuracy, yet reject–repeat analysis (RRA) is still mostly performed manually and depends on subjective radiographer judgment. Inter-operator variability leads to inconsistent quality data and limits workflow optimization. While many studies evaluate deep learning models for disease classification, research using convolutional neural networks (CNNs) to assess image quality and automate RRA remains limited. This study introduces a novel approach by applying CNNs to classify the causes of image rejection in chest radiography: positioning, exposure, artefacts, and accepted images, integrating these outcomes into a quality-management framework. The aim is to develop and evaluate CNN models (visual geometry group (VGG)-16, residual neural network (ResNet)-101, and dense convolutional network (DenseNet)-201) to support automated RRA using digital imaging and communications in medicine (DICOM)-based radiographs. The dataset comprised DICOM chest images annotated by a panel of radiologists, radiographers, and medical physicists into four technical classes. The CNNs were trained using supervised learning with structured training, validation, and testing splits, and evaluated using accuracy, precision, recall, and F1-score. Results show VGG-16 delivers the most stable and accurate performance across categories. Overall, CNNs effectively automate RRA with greater objectivity and consistency, reducing repeat rates, improving workflow efficiency, and strengthening radiology quality-management systems to enhance patient safety.

Hybrid Machine Learning for Early Prediction of At-Risk Students with Imbalanced Data Esti Wijayanti; Widowati Widowati; Catur Edi Widodo
Journal of Applied Data Sciences Vol 7, No 2: May 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v7i2.1368

The phenomenon of student dropout remains a major challenge for higher education institutions because it impacts academic performance and institutional reputation. Identification of students at risk of dropping out is often hampered by data imbalance, where the number of dropouts is far fewer than active students, so conventional prediction models tend to be biased towards the majority class. This study aims to develop an accurate and reliable prediction framework for students at risk of dropping out to detect at-risk students through a hybrid machine learning approach with data balancing techniques. The main contribution of this study is the integration of Support Vector Machine and Extreme Gradient Boosting in a stacked ensemble architecture supported by data balancing optimization techniques. The proposed model leverages the ability of Support Vector Machine to separate complex classification patterns, while Extreme Gradient Boosting improves prediction accuracy through iterative learning and modeling interactions between variables. The problem of data imbalance is addressed through oversampling techniques for the minority class so that the model learning process becomes more balanced. The model framework is tested using a dataset consisting of 3,652 students with academic, socioeconomic, and behavioral variables. Experimental results show that the proposed hybrid model outperforms the single model, with an accuracy rate of 97 percent, a precision rate of 94 percent, and a recall rate of 95 percent. These findings suggest that a combination of complementary machine learning methods, coupled with data optimization, can significantly improve the predictive ability of student dropout. The practical implication of this research is the availability of a robust decision support system for universities in designing timely and targeted interventions. By identifying students at risk of dropping out, institutions can strengthen retention strategies, improve student academic success, and reduce dropout rates more effectively.

Title

Found 6 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 6 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 6 Documents
Search