Fenina Adline Twince Tobing
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Fenina Adline Twince Tobing has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Adhi Kusnadi
Universitas Multimedia Nusantara
3 shared publications - Ivransa Zuhdi Pane
Universitas Multimedia Nusantara
2 shared publications - Toto Haryanto
IPB University
2 shared publications - Cian Ramadhona Hassolthine
Universitas Siber Asia
1 shared publication Ghulam Jilani Waqas
Government College University Lahore, Lahore, PAKISTAN
1 shared publicationJansen Wiratama
Universitas Multimedia Nusantara
1 shared publication
Latest Publications
- Open Access
- 7 July 2026
- Published by Universitas Airlangga in Journal of Information Systems Engineering and Business Intelligence
- Vol. 12 (2) , 320-332
Real-Time Meat Authentication Using Deep Convolutional Neural Network with ResNet-50 Architecture
Background: Many issues arise from meat adulteration with regard to food safety and halal certification. Because pork is so visually similar to beef, removing or identifying either type of product becomes almost impossible without some sort of testing. Unfortunately, for many situations, traditional methods for testing authenticity are typically lengthy, costly, and impractical. Objective: This study utilizes computer vision to authenticate beef, pork, and mixtures of both using a deep learning approach. The method used consisted of preprocessing the photographs and performing various forms of data augmentation to increase the dataset’s representative and generalizing capabilities. Methods: We achieved our highest classification accuracy through transfer learning using ResNet-50 as the primary model used for classifying the different types of meats covered by the study, providing empirical metrics and K-fold cross-validation to support reliable classification results. A comparative evaluation using MobileNetV2, EfficientNetB0, and InceptionV3 was also conducted to validate the effectiveness of the proposed ResNet 50 model. Results: The experiment showed that the proposed ResNet-50 model could classify images into their respective classes with a high level of accuracy (98.33%) and greater than 97% for both precision and recall across all classes, in addition to an average inference time of 0.124 s per image. Furthermore, the 5-fold cross-validation yielded an average classification accuracy of 99.33% ± 0.37%, indicating that the performance was stable across multiple data partitions. The low inference time demonstrates the feasibility of the proposed model for real-time meat authentication applications on mobile devices. Conclusion: Overall, the findings indicate that the proposed approach is effective for meat product verification and that computer vision solutions can support the development of practical food safety and halal verification applications. Keywords: Meat authentication, Deep learning, Convolutional Neural Network, ResNet-50, Real-time classification, Food safety
- 7 July 2026
- Published by Universitas Airlangga in Journal of Information Systems Engineering and Business Intelligence
- Vol. 12 (2) , 320-332
Open AccessReal-Time Meat Authentication Using Deep Convolutional Neural Network with ResNet-50 Architecture
Background: Many issues arise from meat adulteration with regard to food safety and halal certification. Because pork is so visually similar to beef, removing or identifying either type of product becomes almost impossible without some sort of testing. Unfortunately, for many situations, traditional methods for testing authenticity are typically lengthy, costly, and impractical. Objective: This study utilizes computer vision to authenticate beef, pork, and mixtures of both using a deep learning approach. The method used consisted of preprocessing the photographs and performing various forms of data augmentation to increase the dataset’s representative and generalizing capabilities. Methods: We achieved our highest classification accuracy through transfer learning using ResNet-50 as the primary model used for classifying the different types of meats covered by the study, providing empirical metrics and K-fold cross-validation to support reliable classification results. A comparative evaluation using MobileNetV2, EfficientNetB0, and InceptionV3 was also conducted to validate the effectiveness of the proposed ResNet 50 model. Results: The experiment showed that the proposed ResNet-50 model could classify images into their respective classes with a high level of accuracy (98.33%) and greater than 97% for both precision and recall across all classes, in addition to an average inference time of 0.124 s per image. Furthermore, the 5-fold cross-validation yielded an average classification accuracy of 99.33% ± 0.37%, indicating that the performance was stable across multiple data partitions. The low inference time demonstrates the feasibility of the proposed model for real-time meat authentication applications on mobile devices. Conclusion: Overall, the findings indicate that the proposed approach is effective for meat product verification and that computer vision solutions can support the development of practical food safety and halal verification applications. Keywords: Meat authentication, Deep learning, Convolutional Neural Network, ResNet-50, Real-time classification, Food safety
- Open Access
- 1 August 2025
- Published by Institute of Advanced Engineering and Science in Indonesian Journal of Electrical Engineering and Computer Science
- Vol. 39 (2) , 1092-1099
Deepfake detection using convolutional neural networks: a deep learning approach for digital security
The development of artificial intelligence technology, especially deep learning, has facilitated the emergence of increasingly sophisticated deepfake technology. Deepfakes utilize generative adversarial networks (GANs) to manipulate images or videos, making it appear as if someone said or did things that never actually happened. As a result, deepfake detection has become a critical challenge, particularly in the context of the spread of false information and digital crime. The purpose of this research is to create a method for detecting deepfakes using a convolutional neural network (CNN) approach, which has been proven effective in visual pattern recognition. Through training with a dataset of original facial images and deepfakes, the CNN model achieved an accuracy of 81.3% in detecting deepfakes. The evaluation results for metrics such as precision, recall, and F1-score indicated good performance overall, although there is still room for improvement. This study is expected to make a significant contribution to enhancing digital security, especially in detecting visual manipulations based on deepfakes.
- 1 August 2025
- Published by Institute of Advanced Engineering and Science in Indonesian Journal of Electrical Engineering and Computer Science
- Vol. 39 (2) , 1092-1099
Open AccessDeepfake detection using convolutional neural networks: a deep learning approach for digital security
The development of artificial intelligence technology, especially deep learning, has facilitated the emergence of increasingly sophisticated deepfake technology. Deepfakes utilize generative adversarial networks (GANs) to manipulate images or videos, making it appear as if someone said or did things that never actually happened. As a result, deepfake detection has become a critical challenge, particularly in the context of the spread of false information and digital crime. The purpose of this research is to create a method for detecting deepfakes using a convolutional neural network (CNN) approach, which has been proven effective in visual pattern recognition. Through training with a dataset of original facial images and deepfakes, the CNN model achieved an accuracy of 81.3% in detecting deepfakes. The evaluation results for metrics such as precision, recall, and F1-score indicated good performance overall, although there is still room for improvement. This study is expected to make a significant contribution to enhancing digital security, especially in detecting visual manipulations based on deepfakes.
- Open Access
- 30 June 2025
- Published by Universitas Multimedia Nusantara in IJNMT (International Journal of New Media Technology)
- Vol. 12 (1) , 55-60
E-Commerce Product Review Sentiment Analysis: A Comparative Study of Naïve Bayes Classifier and Random Forest Algorithms on Marketplace Platforms
Achieving customer satisfaction and trust is a major challenge for success in the business world. Entrepreneurs must identify problems that arise from reviews given by customers. However, reading and sorting each review is time-consuming and considered inefficient. In order to overcome this, a study was conducted that aims to analyze sentiment on products sold in the Shopee marketplace using the Naïve Bayes Classifier and Random Forest algorithms. The focus of this study is on product reviews from XYZ Store. The main objective of this study is to determine a more accurate and efficient algorithm in classifying review sentiment, which can help companies in marketing strategies and product development. The results of this study can provide insight for companies about consumer responses to marketed products, so that they can be used as a basis for making strategic decisions to improve the quality of services and products. The results of the Random Forest method classification produce superior predictions compared to the Naïve Bayes Classifier method with an accuracy value of 92.5%, precision of 93%, Recall of 92.5% and F1-Score of 90%.
- 30 June 2025
- Published by Universitas Multimedia Nusantara in IJNMT (International Journal of New Media Technology)
- Vol. 12 (1) , 55-60
Open AccessE-Commerce Product Review Sentiment Analysis: A Comparative Study of Naïve Bayes Classifier and Random Forest Algorithms on Marketplace Platforms
Achieving customer satisfaction and trust is a major challenge for success in the business world. Entrepreneurs must identify problems that arise from reviews given by customers. However, reading and sorting each review is time-consuming and considered inefficient. In order to overcome this, a study was conducted that aims to analyze sentiment on products sold in the Shopee marketplace using the Naïve Bayes Classifier and Random Forest algorithms. The focus of this study is on product reviews from XYZ Store. The main objective of this study is to determine a more accurate and efficient algorithm in classifying review sentiment, which can help companies in marketing strategies and product development. The results of this study can provide insight for companies about consumer responses to marketed products, so that they can be used as a basis for making strategic decisions to improve the quality of services and products. The results of the Random Forest method classification produce superior predictions compared to the Naïve Bayes Classifier method with an accuracy value of 92.5%, precision of 93%, Recall of 92.5% and F1-Score of 90%.
- Open Access
- 29 April 2025
- Published by Universitas Multimedia Nusantara in Ultimatics : Jurnal Teknik Informatika
- Vol. 17 (1) , 82-89
Enhancing Intelligent Tutoring Systems through SVM-Based Academic Performance Classification and Rule-Based Question Recommendation
The aims to automatically classify students' academic performance levels using Support Vector Machine (SVM) algorithm and automatically recommend questions based on classification results. Dataset consists of six assignment scores per student, averaging students into three performance levels: Beginner, Intermediate, and Advanced. Before training, the data undergoes preprocessing involving normalization with Standard Scaler and splitting into training and testing sets. The model is trained using Radial Basis Function (RBF) kernel with hyperparameter tuning to optimize its performance. The evaluation results show that the model achieved an accuracy of 91.67%, with a precision of 93.06%, a recall of 91.67%, and an F1-score of 91.89%. The best performance was found in the Intermediate class, the dominant category in the dataset, while performance in the Advanced category was relatively lower due to limited sample size. Following classification, a rule-based recommendation system is used to suggest questions that match the student's predicted level of competence. This approach supports a more adaptive and personalized learning environment. The findings demonstrate that the SVM algorithm effectively supports intelligent learning systems such as the Intelligent Tutoring System (ITS). Future work should include data balancing techniques, expansion of dataset size, and comparative analysis with other algorithms to enhance model generalization.
- 29 April 2025
- Published by Universitas Multimedia Nusantara in Ultimatics : Jurnal Teknik Informatika
- Vol. 17 (1) , 82-89
Open AccessEnhancing Intelligent Tutoring Systems through SVM-Based Academic Performance Classification and Rule-Based Question Recommendation
The aims to automatically classify students' academic performance levels using Support Vector Machine (SVM) algorithm and automatically recommend questions based on classification results. Dataset consists of six assignment scores per student, averaging students into three performance levels: Beginner, Intermediate, and Advanced. Before training, the data undergoes preprocessing involving normalization with Standard Scaler and splitting into training and testing sets. The model is trained using Radial Basis Function (RBF) kernel with hyperparameter tuning to optimize its performance. The evaluation results show that the model achieved an accuracy of 91.67%, with a precision of 93.06%, a recall of 91.67%, and an F1-score of 91.89%. The best performance was found in the Intermediate class, the dominant category in the dataset, while performance in the Advanced category was relatively lower due to limited sample size. Following classification, a rule-based recommendation system is used to suggest questions that match the student's predicted level of competence. This approach supports a more adaptive and personalized learning environment. The findings demonstrate that the SVM algorithm effectively supports intelligent learning systems such as the Intelligent Tutoring System (ITS). Future work should include data balancing techniques, expansion of dataset size, and comparative analysis with other algorithms to enhance model generalization.
- Open Access
- 1 April 2025
- Published by Institute of Advanced Engineering and Science in IAES International Journal of Artificial Intelligence (IJ-AI)
- Vol. 14 (2) , 1056-1066
Enhancing facial recognition accuracy through feature extractions and artificial neural networks
Facial recognition is a biometric system used to identify individuals through faces. Although this technology has many advantages, it still faces several challenges. One of the main challenges is that the level of accuracy has yet to reach its maximum potential. This research aims to improve facial recognition performance by applying the discrete cosine transform (DCT) and Gaussian mixture model (GMM), which are then trained with backward propagation of errors (backpropagation) and convolutional neural networks (CNN). The research results show low DCT and GMM feature extraction accuracy with backpropagation of 4.88%. However, the combination of DCT, GMM, and CNN feature extraction produces an accuracy of up to 98.2% and a training time of 360 seconds on the Olivetti Research Laboratory (ORL) dataset, an accuracy of 98.9% and a training time of 1210 seconds on the Yale dataset, and 100% accuracy and training time 1749 seconds on the Japanese female facial expression (JAFFE) dataset. This improvement is due to the combination of DCT, GMM, and CNN's ability to remove noise and study images accurately. This research is expected to significantly contribute to overcoming accuracy challenges and increasing the flexibility of facial recognition systems in various practical situations, as well as the potential to improve security and reliability in security and biometrics.
- 1 April 2025
- Published by Institute of Advanced Engineering and Science in IAES International Journal of Artificial Intelligence (IJ-AI)
- Vol. 14 (2) , 1056-1066
Open AccessEnhancing facial recognition accuracy through feature extractions and artificial neural networks
Facial recognition is a biometric system used to identify individuals through faces. Although this technology has many advantages, it still faces several challenges. One of the main challenges is that the level of accuracy has yet to reach its maximum potential. This research aims to improve facial recognition performance by applying the discrete cosine transform (DCT) and Gaussian mixture model (GMM), which are then trained with backward propagation of errors (backpropagation) and convolutional neural networks (CNN). The research results show low DCT and GMM feature extraction accuracy with backpropagation of 4.88%. However, the combination of DCT, GMM, and CNN feature extraction produces an accuracy of up to 98.2% and a training time of 360 seconds on the Olivetti Research Laboratory (ORL) dataset, an accuracy of 98.9% and a training time of 1210 seconds on the Yale dataset, and 100% accuracy and training time 1749 seconds on the Japanese female facial expression (JAFFE) dataset. This improvement is due to the combination of DCT, GMM, and CNN's ability to remove noise and study images accurately. This research is expected to significantly contribute to overcoming accuracy challenges and increasing the flexibility of facial recognition systems in various practical situations, as well as the potential to improve security and reliability in security and biometrics.
- Open Access
- 30 June 2024
- Published by Universitas Multimedia Nusantara in IJNMT (International Journal of New Media Technology)
- Vol. 11 (1) , 27-34
Implementation of Support Vector Machine Method for Twitter Sentiment Analysis Related to Cancellation of u-20 World Cup in Indonesia
The cancellation of the U-20 world cup in Indonesia in 2023 has become a hot debate among the Indonesian people because the reasons for the cancellation are still unclear. The number of pro and con opinions uploaded by the Indonesian people on twitter social media makes these opinions can be used as data to assess opinions which are divided into three categories, namely positive, negative and neutral. After being divided into three categories, sentiment analysis will then be carried out using the SVM method and comparing linear, polynomial and rbf kernels to get the best performance of existing kernels in the support vector machine method. By using confusion matrix to measure the performance of the classification, accuracy, precision, recall and f1-score can be assessed. It was found that the 80:20 data ratio had the highest accuracy of the linear, polynomial, rbf kernel and the rbf kernel had better results than the linear and polynomial kernels, namely Accuracy 78.15%, F1-Score, 76.30%, Precision 77.37% and Recall 75.58%. In addition, the data obtained also succeeded in analyzing Indonesian texts that were input externally and categorized into positive, neutral and negative. From the results that have been obtained, the support vector machine method has been successfully implemented in sentiment analysis of the U-20 world cup cancellation in Indonesia in 2023 on twitter social media
- 30 June 2024
- Published by Universitas Multimedia Nusantara in IJNMT (International Journal of New Media Technology)
- Vol. 11 (1) , 27-34
Open AccessImplementation of Support Vector Machine Method for Twitter Sentiment Analysis Related to Cancellation of u-20 World Cup in Indonesia
The cancellation of the U-20 world cup in Indonesia in 2023 has become a hot debate among the Indonesian people because the reasons for the cancellation are still unclear. The number of pro and con opinions uploaded by the Indonesian people on twitter social media makes these opinions can be used as data to assess opinions which are divided into three categories, namely positive, negative and neutral. After being divided into three categories, sentiment analysis will then be carried out using the SVM method and comparing linear, polynomial and rbf kernels to get the best performance of existing kernels in the support vector machine method. By using confusion matrix to measure the performance of the classification, accuracy, precision, recall and f1-score can be assessed. It was found that the 80:20 data ratio had the highest accuracy of the linear, polynomial, rbf kernel and the rbf kernel had better results than the linear and polynomial kernels, namely Accuracy 78.15%, F1-Score, 76.30%, Precision 77.37% and Recall 75.58%. In addition, the data obtained also succeeded in analyzing Indonesian texts that were input externally and categorized into positive, neutral and negative. From the results that have been obtained, the support vector machine method has been successfully implemented in sentiment analysis of the U-20 world cup cancellation in Indonesia in 2023 on twitter social media
- Open Access
- 3 August 2022
- Published by Universitas Multimedia Nusantara in Ultima InfoSys : Jurnal Ilmu Sistem Informasi
- Vol. 13 (1) , 19-28
Analysis and Design of an Web-Based Ticketing Service Helpdesk at Food and Packaging Machinery Company
PT Putra Chandra Sentosa is a company engaged in food and packaging machinery. In the business process that runs in the customer service division, the process of recording service submissions by customer service is carried out using word processing software which has the potential for data accumulation and the search for information on service submissions to be slower. Customer service has difficulty dividing technicians' duties due to service status updates that are not real-time, and they have to wait for technicians to submit a letter of assignment after the visit. In addition to using conventional data processing, customers who complain about service results or technician attitudes make management feel they need feedback reports regarding customer satisfaction. This study aims to design a website that is useful to facilitate the search for service information, perform a balanced division of technician tasks and assist management in seeing customer satisfaction in the form of reports. The Helpdesk Ticketing website in this study has been developed using Website Development Life Cycle (WDLC) model with the Hypertext Preprocessor (PHP) programming language, Unified Modeling Language (UML) notation and MySQL database. The PHP programming language is used in this research because various operating systems can run it via a browser. This research produces a Helpdesk Ticketing website that can answer the needs of the company PT Putra Chandra Sentosa. This research also proves that the website can optimize business processes in recording service submissions compared to using word processing software
- 3 August 2022
- Published by Universitas Multimedia Nusantara in Ultima InfoSys : Jurnal Ilmu Sistem Informasi
- Vol. 13 (1) , 19-28
Open AccessAnalysis and Design of an Web-Based Ticketing Service Helpdesk at Food and Packaging Machinery Company
PT Putra Chandra Sentosa is a company engaged in food and packaging machinery. In the business process that runs in the customer service division, the process of recording service submissions by customer service is carried out using word processing software which has the potential for data accumulation and the search for information on service submissions to be slower. Customer service has difficulty dividing technicians' duties due to service status updates that are not real-time, and they have to wait for technicians to submit a letter of assignment after the visit. In addition to using conventional data processing, customers who complain about service results or technician attitudes make management feel they need feedback reports regarding customer satisfaction. This study aims to design a website that is useful to facilitate the search for service information, perform a balanced division of technician tasks and assist management in seeing customer satisfaction in the form of reports. The Helpdesk Ticketing website in this study has been developed using Website Development Life Cycle (WDLC) model with the Hypertext Preprocessor (PHP) programming language, Unified Modeling Language (UML) notation and MySQL database. The PHP programming language is used in this research because various operating systems can run it via a browser. This research produces a Helpdesk Ticketing website that can answer the needs of the company PT Putra Chandra Sentosa. This research also proves that the website can optimize business processes in recording service submissions compared to using word processing software