Claim Missing Document
Check
Articles

Found 16 Documents
Search

KLASIFIKASI EMOSI MENGGUNAKAN COMPUTER VISION DAN CONVOLUTIONAL NEURAL NETWORKS Nurdiyansah , Andri; Rochmawati, Dwi Robiul
Jurnal Teknologi Komputer dan Informatika Vol 3 No 2 (2025): Jurnal Teknologi Komputer dan Informatika (TEKOMIN)
Publisher : LPPM Politeknik Pajajaran ICB Bandung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59820/tekomin.v3i2.411

Abstract

This research focuses on the development of an emotion classification system utilizing computer vision and Convolutional Neural Networks (CNN). This model was trained on the FER2013 dataset, which contains 35,809 facial images categorized into seven emotions. Metode seperti augmentasi data dan normalisasi piksel diterapkan untuk meningkatkan ketahanan model. The CNN architecture achieved an accuracy of 85%, demonstrating its effectiveness in recognizing emotions such as happiness and anger. This research highlights the potential integration of emotion-aware systems into applications such as human-computer interaction and personalized services, emphasizing technical innovation in AI-based solutions. Keywords: Emotion Classification; Computer Vision; CNN; FER2013; Deep Learning
PREDIKSI CUACA DENGAN JARINGAN SYARAF TIRUAN MENGGUNAKAN PYTHON Rochmawati, Dwi Robiul
Jurnal Teknologi Komputer dan Informatika Vol 2 No 2 (2024): Jurnal Teknologi Komputer dan Informatika (TEKOMIN)
Publisher : LPPM Politeknik Pajajaran ICB Bandung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59820/tekomin.v2i2.228

Abstract

Weather prediction is one of the growing challenges in meteorological science. Accurate prediction methods can provide invaluable information for various sectors, including agriculture, transport, and natural disaster mitigation. One approach used in predicting weather is using artificial neural network (ANN) techniques. JST is a computational model inspired by the structure and function of human biological neural networks. This research aims to implement and evaluate the performance of JST in weather prediction. The data used is historical weather data that includes parameters such as air temperature, humidity, air pressure, and wind direction. The training process is carried out using a suitable learning algorithm to adjust the weights in the JST to produce accurate weather predictions. The results of this study show that a single hidden layer with only two nodes performs slightly better than more complex architectures. In addition, it requires a much shorter training time. In terms of accuracy and efficiency despite using a simpler architecture, the small network almost achieved the same accuracy (around 89%) as the original network. In addition, its training time is also more efficient. So based on these findings, it was decided to continue with the optimized network layout (one hidden layer with 2 nodes) due to the good balance between accuracy and efficiency. This research not only improves the accuracy of weather prediction but also highlights the importance of neural network architecture optimization according to the specific dataset and task.
MANFAAT KECERDASAN BUATAN UNTUK PENDIDIKAN Rochmawati, Dwi Robiul; Arya , Ivan; Zakariyya, Azka
Jurnal Teknologi Komputer dan Informatika Vol 2 No 1 (2023): Jurnal Teknologi Komputer dan Informatika (TEKOMIN)
Publisher : LPPM Politeknik Pajajaran ICB Bandung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59820/tekomin.v2i1.163

Abstract

Pemanfaatan kecerdasan buatan (AI) dalam pendidikan telah menjadi subjek yang menarik perhatian akademisi dan praktisi pendidikan. Dalam konteks ini, penelitian ini bertujuan untuk mengeksplorasi manfaat AI dalam pembelajaran, pengembangan metode pengajaran, serta peningkatan efektivitas secara keseluruhan. Melalui metode penelitian deskriptif dan analisis komprehensif, penelitian ini mengidentifikasi manfaat AI dalam personalisasi pembelajaran, pengajaran adaptif, analisis data pendidikan, serta peningkatan akses global terhadap pendidikan. Namun, penerapan AI juga menimbulkan sejumlah tantangan dan kekhawatiran etika, seperti bias algoritma, privasi dan keamanan data, ketergantungan teknologi, dan kurangnya pelatihan guru. Oleh karena itu, penelitian ini memberikan saran-saran untuk mengelola tantangan tersebut secara bijaksana, termasuk pengelolaan bias algoritma, perlindungan privasi data, keterlibatan manusia yang cukup, transparansi algoritma, pelatihan guru yang memadai, dan pertimbangan etika dalam pengambilan keputusan karier. Dengan pendekatan yang hati-hati terhadap tantangan dan saran-saran ini, integrasi AI dalam pendidikan dapat memberikan manfaat maksimal sambil menjaga keadilan, privasi, dan keterlibatan manusia yang penting dalam proses pendidikan.
Deep Learning-Based ResNet-50 Transfer Learning Approaches for Pneumonia Detection from Chest X-Ray Images: With and Without Fine-Tuning Rochmawati, Dwi Robiul; Maryani, Lidya
Indonesian Journal of Health Research and Development Vol. 3 No. 3 (2025): Indonesian Journal of Health Research and Development
Publisher : CV Media Inti Teknologi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.58723/ijhrd.v3i3.507

Abstract

Background: Pneumonia remains one of the leading causes of morbidity and mortality worldwide, particularly among children and older adults in low-resource settings. Diagnosis based on chest X-ray interpretation often depends on radiologist expertise, which may be limited in availability and prone to subjectivity. Deep learning offers a promising alternative to improve diagnostic efficiency and consistency.Aims: This study aims to evaluate the effectiveness of the ResNet-50 architecture for pneumonia detection using chest X-ray images by comparing transfer learning with frozen layers and partial fine-tuning strategies.Methods: A total of 5,856 chest X-ray images were obtained from a public dataset and divided into training, validation, and testing sets using stratified sampling. Data preprocessing included resizing, normalization, and augmentation. Two models were developed: (1) a frozen ResNet-50 model, where all convolutional layers were fixed, and (2) a fine-tuned ResNet-50 model, where the final convolutional layers were retrained. Performance was evaluated using accuracy, precision, recall, F1-score, and area under the ROC curve (AUC). Statistical tests were conducted to assess performance differences between the two models.Results: The frozen model achieved an accuracy of 62.50% and an AUC of 0.4819, indicating weak classification performance. In contrast, the fine-tuned model demonstrated substantially higher accuracy of 85.90%, F1-score of 0.8967, and AUC of 0.9510, showing strong discriminative capability. Statistical analysis confirmed that the performance improvement in accuracy was significant.Conclusion: Fine-tuning significantly enhances the applicability of ResNet-50 for pneumonia detection. Without feature adaptation, pretrained models struggle to generalize to medical imaging domains. Fine-tuned transfer learning provides a more reliable framework for developing computer-aided diagnostic systems, particularly in clinical environments with limited expert availability.
Classification and Interpretability of Employee Burnout Using Linear Discriminant Analysis Rochmawati, Dwi Robiul; Muhammad Al Adib; Diyo Mollana Fazri; Bill Raj; Romi Antoni; Rahmad Santoso; Wahyu Saptha Negoro
Jurnal Komputer Teknologi Informasi Sistem Komputer (JUKTISI) Vol. 4 No. 3 (2026): Februari 2026
Publisher : LKP KARYA PRIMA KURSUS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62712/juktisi.v4i3.811

Abstract

Employee burnout has become a critical challenge in modern organizations due to its negative impact on employees’ mental well-being, work performance, and organizational sustainability. In many workplaces, burnout identification still relies on subjective assessments and retrospective surveys, limiting the effectiveness of early intervention strategies. This study aims to develop an employee burnout risk classification model that achieves high predictive performance while maintaining strong interpretability. Linear Discriminant Analysis (LDA) is employed as the primary method because of its ability to separate classes optimally and provide explicit discriminant coefficients for explanatory analysis. The study utilizes a secondary dataset from the Mental Health in Workplace Survey, consisting of 3,000 employee records and 15 variables related to job characteristics, psychosocial factors, and individual conditions. The dataset is divided into training and testing sets with an 80:20 ratio. Experimental results show that the LDA model achieves an accuracy of 96.17%, with a precision of 89.50%, recall of 100%, F1-score of 94.46%, and an AUC value of 0.9988, indicating excellent classification capability. Further analysis of discriminant coefficients reveals that individual burnout indicators, job roles, work–life balance, and career growth opportunities are the most influential factors in determining burnout risk. These findings demonstrate that LDA offers an effective and interpretable approach for early burnout detection and supports evidence-based decision-making for human resource management.
Analisis SelectKBest pada Klasifikasi Trafik VPN Menggunakan Random Forest dan SVM Nurdiansyah, Andri; R, Dwi Robiul; Sururi, Sururi; Sujana, Nana
TIN: Terapan Informatika Nusantara Vol 6 No 8 (2026): January 2026
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/tin.v6i8.9136

Abstract

The increasing use of Virtual Private Networks (VPNs) in modern networks poses significant challenges for network monitoring and traffic management, particularly in accurately and efficiently distinguishing VPN and non-VPN traffic. This study aims to analyze the effectiveness of the SelectKBest feature selection method in improving VPN traffic classification performance using Random Forest and Support Vector Machine (SVM) algorithms. The dataset used in this study is the CIC VPN-NonVPN Traffic Dataset provided by the Canadian Institute for Cybersecurity (CIC), which is widely recognized as a standard benchmark in network security research. Feature selection was performed using SelectKBest with the ANOVA (f_classif) scoring function, reducing the original feature set to 15 most relevant features. Experimental results show that the Random Forest classifier achieved a test accuracy of 84.94%, along with high F1-score and ROC-AUC values, and an average cross-validation accuracy of 95.18% with low variance. In contrast, the SVM model demonstrated relatively poor performance, with a test accuracy of approximately 62%, indicating its limitation in capturing the complex patterns of network traffic data. Further analysis using ROC curves, Precision–Recall curves, confusion matrices, and learning curves confirms that Random Forest exhibits superior generalization capability compared to SVM. These findings indicate that the combination of SelectKBest and Random Forest not only delivers high classification performance but also improves computational efficiency through feature dimensionality reduction, making it suitable for large-scale VPN traffic classification scenarios.