Imam Tahyudin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Imam Tahyudin has a total of 91 co-authors. Here are the top co-authors by number of shared publications:
Hidetaka Nambo
Kanazawa University
10 shared publications- Dhanar Intan Surya Saputra
Department of Information Systems, Universitas Amikom Purwokerto
4 shared publications Ades Tikaningsih
3 shared publicationsDhanar Intan Surya Saputra
STMIK AMIKOM Purwokerto
3 shared publications- Hidetaka Nambo
Graduate School of Natural Science and Technology, Kanazawa University
3 shared publications - Taqwa Hariguna
Department of Information System and Magister Computer Sciences, Universitas Amikom Purwokerto, Purwokerto Utara, 53127, Indonesia
3 shared publications
Latest Publications
- Open Access
- 30 July 2026
- Published by Universitas Muhammadiyah Mataram in JMM (Jurnal Masyarakat Mandiri)
- Vol. 10 (4) , 4986-4997
TRANSNATIONAL ACADEMIC-COMMUNITY PARTNERSHIPS: FACILITATING FINANCIAL CAPABILITY AND REMOTE HIGHER EDUCATION FOR INDONESIAN OVERSEAS WORKERS IN JAPAN
Abstrak: Banyak Pekerja Migran Indonesia (PMI) di Jepang menghadapi masalah keterbatasan waktu belajar dan kerentanan terhadap investasi ilegal. Kegiatan pengabdian masyarakat ini bertujuan meningkatkan kapabilitas keuangan dan akses pendidikan jarak jauh bagi PMI di Jepang melalui kolaborasi antara Universitas Amikom Purwokerto dan mitra Lakpesdam PCINU Jepang. Program dilaksanakan melalui webinar interaktif yang diikuti oleh 46 peserta secara daring. Evaluasi dilakukan secara kuantitatif menggunakan kuesioner pre-test dan post-test dengan 15 butir pertanyaan. Indikator keberhasilan diukur dari peningkatan pemahaman kognitif. Hasil menunjukkan peningkatan rata-rata pemahaman peserta sebesar 17,1%. Kemitraan ini menunjukkan bahwa organisasi diaspora berfungsi efektif sebagai penghubung kepercayaan untuk sosialisasi akademik, meskipun peserta masih membutuhkan pendampingan berkelanjutan untuk membagi waktu antara bekerja dan belajar daring. Many Indonesian Overseas Workers (IOWs) in Japan face learning barriers and exposure to predatory financial schemes. This community service project aimed to enhance financial capability and Remote Higher Education (RHE) accessibility for IOW in Japan through a collaborative partnership between Universitas Amikom Purwokerto and Lakpesdam PCINU Japan. Utilizing an interactive online webinar, the initiative engaged 46 participants. The educational impact was assessed quantitatively using pre-test and post-test questionnaires containing 15 items. Success indicators were measured by improvements in cognitive comprehension. Results showed a 17.1% increase in overall comprehension. The findings indicate that diaspora networks serve as effective trust brokers for academic socialization. However, sustained mentoring support remains crucial to assist working students in balancing their professional duties and online learning.
- 30 July 2026
- Published by Universitas Muhammadiyah Mataram in JMM (Jurnal Masyarakat Mandiri)
- Vol. 10 (4) , 4986-4997
Open AccessTRANSNATIONAL ACADEMIC-COMMUNITY PARTNERSHIPS: FACILITATING FINANCIAL CAPABILITY AND REMOTE HIGHER EDUCATION FOR INDONESIAN OVERSEAS WORKERS IN JAPAN
Abstrak: Banyak Pekerja Migran Indonesia (PMI) di Jepang menghadapi masalah keterbatasan waktu belajar dan kerentanan terhadap investasi ilegal. Kegiatan pengabdian masyarakat ini bertujuan meningkatkan kapabilitas keuangan dan akses pendidikan jarak jauh bagi PMI di Jepang melalui kolaborasi antara Universitas Amikom Purwokerto dan mitra Lakpesdam PCINU Jepang. Program dilaksanakan melalui webinar interaktif yang diikuti oleh 46 peserta secara daring. Evaluasi dilakukan secara kuantitatif menggunakan kuesioner pre-test dan post-test dengan 15 butir pertanyaan. Indikator keberhasilan diukur dari peningkatan pemahaman kognitif. Hasil menunjukkan peningkatan rata-rata pemahaman peserta sebesar 17,1%. Kemitraan ini menunjukkan bahwa organisasi diaspora berfungsi efektif sebagai penghubung kepercayaan untuk sosialisasi akademik, meskipun peserta masih membutuhkan pendampingan berkelanjutan untuk membagi waktu antara bekerja dan belajar daring. Many Indonesian Overseas Workers (IOWs) in Japan face learning barriers and exposure to predatory financial schemes. This community service project aimed to enhance financial capability and Remote Higher Education (RHE) accessibility for IOW in Japan through a collaborative partnership between Universitas Amikom Purwokerto and Lakpesdam PCINU Japan. Utilizing an interactive online webinar, the initiative engaged 46 participants. The educational impact was assessed quantitatively using pre-test and post-test questionnaires containing 15 items. Success indicators were measured by improvements in cognitive comprehension. Results showed a 17.1% increase in overall comprehension. The findings indicate that diaspora networks serve as effective trust brokers for academic socialization. However, sustained mentoring support remains crucial to assist working students in balancing their professional duties and online learning.
- Open Access
- 30 December 2025
- Published by Politeknik Negeri Cilacap in Journal of Innovation Information Technology and Application (JINITA)
- Vol. 7 (2) , 406-412
Automated Esophagitis Detection from Endoscopy Using Deep Learning
Gastro-esophageal reflux disease (GERD) is a widespread condition that often leads to severe complications, including esophagitis, which significantly affects patient health and quality of life. While endoscopy is the gold standard for diagnosing esophagitis, its reliance on specialized equipment and trained professionals can limit accessibility and efficiency. This study introduces an innovative approach to diagnosing esophagitis by leveraging Convolutional Neural Networks (CNN) for automated classification of endoscopic images. By utilizing the Kvasir dataset, which includes a comprehensive collection of gastrointestinal endoscopy images, the model is trained to distinguish between esophagitis and normal-Z-line conditions with remarkable accuracy. The CNN model achieved outstanding results, with an accuracy of 96.04%, precision of 98.94%, recall of 93.00%, and an F1-score of 95.88%, demonstrating its potential to outperform traditional diagnostic methods. These findings underscore the ability of CNN to not only enhance diagnostic precision but also to reduce human error, making the process faster, more reliable, and more accessible. This research contributes to the growing body of work in medical image analysis, suggesting that CNN-based models can transform clinical practices by supporting timely, accurate diagnoses while alleviating the burden on medical professionals. The integration of deep learning in this domain holds the promise of advancing healthcare accessibility and efficiency globally
- 30 December 2025
- Published by Politeknik Negeri Cilacap in Journal of Innovation Information Technology and Application (JINITA)
- Vol. 7 (2) , 406-412
Open AccessAutomated Esophagitis Detection from Endoscopy Using Deep Learning
Gastro-esophageal reflux disease (GERD) is a widespread condition that often leads to severe complications, including esophagitis, which significantly affects patient health and quality of life. While endoscopy is the gold standard for diagnosing esophagitis, its reliance on specialized equipment and trained professionals can limit accessibility and efficiency. This study introduces an innovative approach to diagnosing esophagitis by leveraging Convolutional Neural Networks (CNN) for automated classification of endoscopic images. By utilizing the Kvasir dataset, which includes a comprehensive collection of gastrointestinal endoscopy images, the model is trained to distinguish between esophagitis and normal-Z-line conditions with remarkable accuracy. The CNN model achieved outstanding results, with an accuracy of 96.04%, precision of 98.94%, recall of 93.00%, and an F1-score of 95.88%, demonstrating its potential to outperform traditional diagnostic methods. These findings underscore the ability of CNN to not only enhance diagnostic precision but also to reduce human error, making the process faster, more reliable, and more accessible. This research contributes to the growing body of work in medical image analysis, suggesting that CNN-based models can transform clinical practices by supporting timely, accurate diagnoses while alleviating the burden on medical professionals. The integration of deep learning in this domain holds the promise of advancing healthcare accessibility and efficiency globally
- Open Access
- 24 December 2025
- Published by Infinite Corporation in Jurnal Pendidikan dan Teknologi Indonesia
- Vol. 5 (12) , 3682-3692
Evaluasi Ensemble Learning untuk Prediksi Nilai Matematika Siswa Sekolah Menengah
Prediksi dini performa matematika siswa sekolah menengah sangat penting untuk merancang intervensi pendidikan yang lebih adaptif dan efektif sebelum ujian akhir resmi dilaksanakan. Penelitian ini bertujuan untuk mengevaluasi kinerja tiga model machine learning Random Forest (RF), Gradient Boosting Regressor (GBR), dan Multi-Layer Perceptron (MLP) dalam memprediksi nilai matematika siswa di Indonesia, serta mendokumentasikan proses tuning hyperparameter secara sistematis untuk setiap model. Dataset yang digunakan terdiri dari skor matematika, membaca, menulis, serta variabel demografis meliputi jenis kelamin, latar belakang pendidikan orang tua, jenis layanan makan, dan keikutsertaan kursus persiapan. Proses tuning hyperparameter untuk RF dan GBR dilakukan menggunakan RandomizedSearchCV dengan 5-fold cross-validation, menguji rentang nilai untuk jumlah estimator, kedalaman maksimum pohon, dan laju pembelajaran (learning rate). Sedangkan pada Multi-Layer Perceptron, GridSearchCV diterapkan dengan variasi arsitektur hidden_layer_sizes, laju pembelajaran awal (learning_rate_init), dan faktor regularisasi (alpha) pada 5-fold CV. Model diukur menggunakan Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), dan koefisien determinasi (R²). Hasil eksperimen menunjukkan bahwa GBR memberikan performa terbaik dengan MAE sebesar 11,61 poin, RMSE 15,23 poin, dan R² 0,10. Random Forest menempati urutan kedua (MAE 12,34; RMSE 16,05; R² 0,64), diikuti MLP (MAE 13,10; RMSE 17,20; R² 0,60). Analisis feature importance mengungkap bahwa skor membaca dan menulis bersama-sama menyumbang lebih dari 60 % kontribusi prediksi, sedangkan faktor demografis seperti latar belakang pendidikan orang tua dan keikutsertaan kursus berperan sekunder namun tetap signifikan. Temuan ini mengindikasikan bahwa model ensemble learning tidak hanya unggul dalam akurasi prediksi, tetapi juga memberikan wawasan mendalam tentang variabel kunci yang memengaruhi performa matematika siswa. Implementasi model ini memungkinkan guru dan pihak sekolah untuk mengidentifikasi siswa yang berisiko rendah secara lebih cepat, merancang program remedial atau pengayaan yang tepat sasaran, serta memanfaatkan sumber daya pendidikan secara lebih efisien. Untuk penelitian lanjutan, disarankan penambahan variabel perilaku siswa seperti durasi belajar mandiri dan kehadiran serta eksplorasi model sekuensial (RNN/Transformer) untuk menangkap dinamika pembelajaran dari waktu ke waktu.
- 24 December 2025
- Published by Infinite Corporation in Jurnal Pendidikan dan Teknologi Indonesia
- Vol. 5 (12) , 3682-3692
Open AccessEvaluasi Ensemble Learning untuk Prediksi Nilai Matematika Siswa Sekolah Menengah
Prediksi dini performa matematika siswa sekolah menengah sangat penting untuk merancang intervensi pendidikan yang lebih adaptif dan efektif sebelum ujian akhir resmi dilaksanakan. Penelitian ini bertujuan untuk mengevaluasi kinerja tiga model machine learning Random Forest (RF), Gradient Boosting Regressor (GBR), dan Multi-Layer Perceptron (MLP) dalam memprediksi nilai matematika siswa di Indonesia, serta mendokumentasikan proses tuning hyperparameter secara sistematis untuk setiap model. Dataset yang digunakan terdiri dari skor matematika, membaca, menulis, serta variabel demografis meliputi jenis kelamin, latar belakang pendidikan orang tua, jenis layanan makan, dan keikutsertaan kursus persiapan. Proses tuning hyperparameter untuk RF dan GBR dilakukan menggunakan RandomizedSearchCV dengan 5-fold cross-validation, menguji rentang nilai untuk jumlah estimator, kedalaman maksimum pohon, dan laju pembelajaran (learning rate). Sedangkan pada Multi-Layer Perceptron, GridSearchCV diterapkan dengan variasi arsitektur hidden_layer_sizes, laju pembelajaran awal (learning_rate_init), dan faktor regularisasi (alpha) pada 5-fold CV. Model diukur menggunakan Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), dan koefisien determinasi (R²). Hasil eksperimen menunjukkan bahwa GBR memberikan performa terbaik dengan MAE sebesar 11,61 poin, RMSE 15,23 poin, dan R² 0,10. Random Forest menempati urutan kedua (MAE 12,34; RMSE 16,05; R² 0,64), diikuti MLP (MAE 13,10; RMSE 17,20; R² 0,60). Analisis feature importance mengungkap bahwa skor membaca dan menulis bersama-sama menyumbang lebih dari 60 % kontribusi prediksi, sedangkan faktor demografis seperti latar belakang pendidikan orang tua dan keikutsertaan kursus berperan sekunder namun tetap signifikan. Temuan ini mengindikasikan bahwa model ensemble learning tidak hanya unggul dalam akurasi prediksi, tetapi juga memberikan wawasan mendalam tentang variabel kunci yang memengaruhi performa matematika siswa. Implementasi model ini memungkinkan guru dan pihak sekolah untuk mengidentifikasi siswa yang berisiko rendah secara lebih cepat, merancang program remedial atau pengayaan yang tepat sasaran, serta memanfaatkan sumber daya pendidikan secara lebih efisien. Untuk penelitian lanjutan, disarankan penambahan variabel perilaku siswa seperti durasi belajar mandiri dan kehadiran serta eksplorasi model sekuensial (RNN/Transformer) untuk menangkap dinamika pembelajaran dari waktu ke waktu.
- Open Access
- 24 December 2025
- Published by Infinite Corporation in Jurnal Pendidikan dan Teknologi Indonesia
- Vol. 5 (12) , 3644-3657
Transformasi Portal Data Pemerintah di Indonesia dengan Large Language Model dan Retrieval-Augmented Generation: Tinjauan Pustaka Sistematis
Integrasi kecerdasan buatan (Artificial Intelligence/AI) seperti Large Language Model (LLM) dan Retrieval-Augmented Generation (RAG) berpotensi mentransformasi portal data pemerintah, namun implementasinya terhambat oleh kurangnya tinjauan sistematis dan kerangka evaluasi yang spesifik. Penelitian ini bertujuan untuk mengidentifikasi, mengevaluasi, dan mensintesis literatur terkini mengenai metodologi, keberhasilan, dan tantangan integrasi teknologi tersebut melalui tinjauan pustaka sistematis. Metode ini diterapkan dengan pencarian terstruktur pada basis data Google Scholar, Scopus, dan IEEE Xplore, diikuti proses penyaringan bertahap. Hasil tinjauan menunjukkan bahwa teknologi AI terbukti efektif meningkatkan komunikasi pemerintah-warga, efisiensi layanan, dan akurasi pengambilan data, di mana penyesuaian model menjadi faktor penting. Namun, implementasinya masih menghadapi tantangan signifikan terkait tata kelola, kualitas data, dan masalah etis. Hasil penelitian ini menekankan pentingnya pengembangan kerangka kerja tata kelola yang komprehensif untuk memastikan penerapan AI yang akuntabel dan selaras dengan kepentingan publik.
- 24 December 2025
- Published by Infinite Corporation in Jurnal Pendidikan dan Teknologi Indonesia
- Vol. 5 (12) , 3644-3657
Open AccessTransformasi Portal Data Pemerintah di Indonesia dengan Large Language Model dan Retrieval-Augmented Generation: Tinjauan Pustaka Sistematis
Integrasi kecerdasan buatan (Artificial Intelligence/AI) seperti Large Language Model (LLM) dan Retrieval-Augmented Generation (RAG) berpotensi mentransformasi portal data pemerintah, namun implementasinya terhambat oleh kurangnya tinjauan sistematis dan kerangka evaluasi yang spesifik. Penelitian ini bertujuan untuk mengidentifikasi, mengevaluasi, dan mensintesis literatur terkini mengenai metodologi, keberhasilan, dan tantangan integrasi teknologi tersebut melalui tinjauan pustaka sistematis. Metode ini diterapkan dengan pencarian terstruktur pada basis data Google Scholar, Scopus, dan IEEE Xplore, diikuti proses penyaringan bertahap. Hasil tinjauan menunjukkan bahwa teknologi AI terbukti efektif meningkatkan komunikasi pemerintah-warga, efisiensi layanan, dan akurasi pengambilan data, di mana penyesuaian model menjadi faktor penting. Namun, implementasinya masih menghadapi tantangan signifikan terkait tata kelola, kualitas data, dan masalah etis. Hasil penelitian ini menekankan pentingnya pengembangan kerangka kerja tata kelola yang komprehensif untuk memastikan penerapan AI yang akuntabel dan selaras dengan kepentingan publik.
- 15 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 289-294
Adaptive Traffic Signal Control: A Comprehensive Bibliometric Analysis
Adaptive Traffic Signal Control (ATSC) has emerged as a key solution to address urban congestion by dynamically adjusting signal timing based on real-time traffic conditions. Despite growing academic interest, systematic mapping of the intellectual structure, thematic evolution, and research gaps in this field remains limited. This study presents a comprehensive bibliometric analysis of 1,287 scholarly articles from Scopus (1973-2024) to map the global landscape of ATSC. Using Bibliometrix (R) and VOSviewer, we performed performance analysis (publication trends, countries, institutions, journals) and scholarly mapping (keyword cooccurrence, thematic clustering). Results show exponential growth since 2015, with dominance from the United States (406 publications) and China (348). Four main thematic clusters were identified: (1) Reinforcement Learning and Multi-Agent Systems, (2) Connected and Autonomous Vehicles (CAVs) as Data Sources, (3) Real-Time Simulation and Field Evaluation, and (4) Sustainability and Resilience in Urban Mobility. This study also reveals future research directions such as federated learning for data privacy, fairness in signal prioritization, and integration with climate adaptation strategies.
- 15 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 289-294
Adaptive Traffic Signal Control: A Comprehensive Bibliometric Analysis
Adaptive Traffic Signal Control (ATSC) has emerged as a key solution to address urban congestion by dynamically adjusting signal timing based on real-time traffic conditions. Despite growing academic interest, systematic mapping of the intellectual structure, thematic evolution, and research gaps in this field remains limited. This study presents a comprehensive bibliometric analysis of 1,287 scholarly articles from Scopus (1973-2024) to map the global landscape of ATSC. Using Bibliometrix (R) and VOSviewer, we performed performance analysis (publication trends, countries, institutions, journals) and scholarly mapping (keyword cooccurrence, thematic clustering). Results show exponential growth since 2015, with dominance from the United States (406 publications) and China (348). Four main thematic clusters were identified: (1) Reinforcement Learning and Multi-Agent Systems, (2) Connected and Autonomous Vehicles (CAVs) as Data Sources, (3) Real-Time Simulation and Field Evaluation, and (4) Sustainability and Resilience in Urban Mobility. This study also reveals future research directions such as federated learning for data privacy, fairness in signal prioritization, and integration with climate adaptation strategies.
- Open Access
- 4 June 2025
- Published by Politeknik Negeri Batam in JOURNAL OF APPLIED INFORMATICS AND COMPUTING
- Vol. 9 (3) , 707-715
Comparative Analysis of VGG16 and ResNet50 Model Performence in Cardiac ECG Image Classification
This study systematically evaluates and compares the effectiveness of two deep learning architectures, VGG16 and ResNet50, in automating electrocardiogram (ECG) image classification for cardiac condition diagnosis. The dataset was obtained from a public source and consists of 2,898 color ECG images converted from raw signals, categorized into four classes: Abnormal Heartbeat, Myocardial Infarction, Normal Individual, and History of Heart Attack. The data underwent preprocessing steps including resizing to 224×224 pixels, pixel normalization to a 0–1 range, label encoding, one-hot encoding, and an 80:20 split for training and testing. Transfer learning was applied using feature representations from the VGG16 and ResNet50 models, employing the Adam optimizer and categorical cross-entropy loss function. To enhance training efficiency and prevent overfitting, early stopping was implemented based on validation loss performance. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics. The results showed that VGG16 achieved 95% accuracy with a loss of 0.1522, precision of 95%, recall of 94%, and F1-score of 94%. In contrast, ResNet50 attained 81% accuracy with a loss of 0.5730, precision of 82%, recall of 79%, and F1-score of 80%. These findings indicate that, within the context of this study, VGG16 consistently outperformed ResNet50 across all evaluation metrics in the ECG image classification task. Therefore, the application of transfer learning using the VGG16 model demonstrates strong potential as an effective approach for AI-based ECG image classification systems.
- 4 June 2025
- Published by Politeknik Negeri Batam in JOURNAL OF APPLIED INFORMATICS AND COMPUTING
- Vol. 9 (3) , 707-715
Open AccessComparative Analysis of VGG16 and ResNet50 Model Performence in Cardiac ECG Image Classification
This study systematically evaluates and compares the effectiveness of two deep learning architectures, VGG16 and ResNet50, in automating electrocardiogram (ECG) image classification for cardiac condition diagnosis. The dataset was obtained from a public source and consists of 2,898 color ECG images converted from raw signals, categorized into four classes: Abnormal Heartbeat, Myocardial Infarction, Normal Individual, and History of Heart Attack. The data underwent preprocessing steps including resizing to 224×224 pixels, pixel normalization to a 0–1 range, label encoding, one-hot encoding, and an 80:20 split for training and testing. Transfer learning was applied using feature representations from the VGG16 and ResNet50 models, employing the Adam optimizer and categorical cross-entropy loss function. To enhance training efficiency and prevent overfitting, early stopping was implemented based on validation loss performance. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics. The results showed that VGG16 achieved 95% accuracy with a loss of 0.1522, precision of 95%, recall of 94%, and F1-score of 94%. In contrast, ResNet50 attained 81% accuracy with a loss of 0.5730, precision of 82%, recall of 79%, and F1-score of 80%. These findings indicate that, within the context of this study, VGG16 consistently outperformed ResNet50 across all evaluation metrics in the ECG image classification task. Therefore, the application of transfer learning using the VGG16 model demonstrates strong potential as an effective approach for AI-based ECG image classification systems.
- Open Access
- 8 April 2025
- Published by Universitas Hamzanwadi in EDUMATIC: Jurnal Pendidikan Informatika
- Vol. 9 (1) , 275-283
Analisis Sentimen Ulasan Co-Pilot Google Play dengan SVM, Neural Network, dan Decision Tree
Sentiment analysis is a technique used to understand user opinions through product or service reviews. The purpose of this research is to compare three classification methods, namely Support Vector Machine (SVM), Neural Network (NN), and Decision Tree (DT) in analyzing the sentiment of users of the Indonesian-language Microsoft Co-Pilot application taken from the Google Play Store. The dataset consists of 20,000 reviews, which first went through preprocessing such as normalization, tokenization, stopwords removal, and stemming. The three methods we used in this study have a Multilayer Perceptron (MLP) architecture with three hidden layers and a ReLU activation function, as well as a dropout regularization technique to avoid overfitting. Model evaluation was conducted using accuracy, precision, recall, and F1-score, with the results showing that NN achieved the highest accuracy of 95.5%, followed by SVM with 95.4% and DT with 92.1%. The advantage of the NN method lies in its ability to recognize more complex patterns in Indonesian, especially in handling informal text and code-mixing. This research contributes to the development of Artificial Intelligence (AI)-based applications by providing insights into the effectiveness of classification methods in Indonesian sentiment analysis, which is important for improving service quality and the development of NLP technology in Indonesia. The practical implications of this research can be used in the development of AI-based applications that are more responsive to user sentiment.
- 8 April 2025
- Published by Universitas Hamzanwadi in EDUMATIC: Jurnal Pendidikan Informatika
- Vol. 9 (1) , 275-283
Open AccessAnalisis Sentimen Ulasan Co-Pilot Google Play dengan SVM, Neural Network, dan Decision Tree
Sentiment analysis is a technique used to understand user opinions through product or service reviews. The purpose of this research is to compare three classification methods, namely Support Vector Machine (SVM), Neural Network (NN), and Decision Tree (DT) in analyzing the sentiment of users of the Indonesian-language Microsoft Co-Pilot application taken from the Google Play Store. The dataset consists of 20,000 reviews, which first went through preprocessing such as normalization, tokenization, stopwords removal, and stemming. The three methods we used in this study have a Multilayer Perceptron (MLP) architecture with three hidden layers and a ReLU activation function, as well as a dropout regularization technique to avoid overfitting. Model evaluation was conducted using accuracy, precision, recall, and F1-score, with the results showing that NN achieved the highest accuracy of 95.5%, followed by SVM with 95.4% and DT with 92.1%. The advantage of the NN method lies in its ability to recognize more complex patterns in Indonesian, especially in handling informal text and code-mixing. This research contributes to the development of Artificial Intelligence (AI)-based applications by providing insights into the effectiveness of classification methods in Indonesian sentiment analysis, which is important for improving service quality and the development of NLP technology in Indonesia. The practical implications of this research can be used in the development of AI-based applications that are more responsive to user sentiment.
- Open Access
- 26 February 2025
- Published by IntSys Research in Techno.Com
- Vol. 24 (1) , 41-55
Perbandingan Algoritma CNN, LSTM, FNN untuk Diagnosa Fibrosis Hati dengan Citra Medis
Fibrosis hati merupakan kondisi yang berpotensi berkembang menjadi sirosis atau kanker hati jika tidak terdiagnosis dengan tepat. Prosedur biopsi hati yang invasif sering digunakan dalam diagnosis, namun memiliki risiko dan keterbatasan biaya. Penelitian ini bertujuan untuk membandingkan kinerja model deep learning, yaitu Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), dan Feedforward Neural Network (FNN) dalam klasifikasi fibrosis hati menggunakan citra medis. Metode yang digunakan adalah evaluasi kinerja model berdasarkan metrik akurasi, presisi, recall, F1-score, dan loss pada dataset citra medis fibrosis hati. Hasil penelitian menunjukkan bahwa CNN memberikan kinerja terbaik dengan akurasi 98%, diikuti oleh LSTM dengan akurasi 97%, dan FNN dengan akurasi 80%. CNN unggul karena kemampuannya dalam mengekstraksi fitur spasial secara otomatis dari citra medis, sementara LSTM lebih cocok untuk data sekuensial dan FNN terbatas dalam menangani data citra kompleks. Penelitian ini menyimpulkan bahwa CNN lebih efektif dalam klasifikasi fibrosis hati dan dapat menjadi alternatif non-invasif yang lebih efisien dibandingkan metode konvensional seperti biopsi. Teknologi ini berpotensi mempercepat diagnosis fibrosis hati dengan akurasi tinggi dan tanpa risiko komplikasi invasif. Kata kunci: Fibrosis hati, CNN, LSTM, FNN, klasifikasi citra medis.
- 26 February 2025
- Published by IntSys Research in Techno.Com
- Vol. 24 (1) , 41-55
Open AccessPerbandingan Algoritma CNN, LSTM, FNN untuk Diagnosa Fibrosis Hati dengan Citra Medis
Fibrosis hati merupakan kondisi yang berpotensi berkembang menjadi sirosis atau kanker hati jika tidak terdiagnosis dengan tepat. Prosedur biopsi hati yang invasif sering digunakan dalam diagnosis, namun memiliki risiko dan keterbatasan biaya. Penelitian ini bertujuan untuk membandingkan kinerja model deep learning, yaitu Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), dan Feedforward Neural Network (FNN) dalam klasifikasi fibrosis hati menggunakan citra medis. Metode yang digunakan adalah evaluasi kinerja model berdasarkan metrik akurasi, presisi, recall, F1-score, dan loss pada dataset citra medis fibrosis hati. Hasil penelitian menunjukkan bahwa CNN memberikan kinerja terbaik dengan akurasi 98%, diikuti oleh LSTM dengan akurasi 97%, dan FNN dengan akurasi 80%. CNN unggul karena kemampuannya dalam mengekstraksi fitur spasial secara otomatis dari citra medis, sementara LSTM lebih cocok untuk data sekuensial dan FNN terbatas dalam menangani data citra kompleks. Penelitian ini menyimpulkan bahwa CNN lebih efektif dalam klasifikasi fibrosis hati dan dapat menjadi alternatif non-invasif yang lebih efisien dibandingkan metode konvensional seperti biopsi. Teknologi ini berpotensi mempercepat diagnosis fibrosis hati dengan akurasi tinggi dan tanpa risiko komplikasi invasif. Kata kunci: Fibrosis hati, CNN, LSTM, FNN, klasifikasi citra medis.
- Open Access
- 25 January 2025
- Published by Bright Publisher in Journal of Applied Data Sciences
- Vol. 6 (1) , 788-799
High-Accuracy Stroke Detection System Using a CBAM-ResNet18 Deep Learning Model on Brain CT Images
High-Accuracy Stroke Detection System Using a CBAM-ResNet18 Deep Learning Model on Brain CT Images
- 25 January 2025
- Published by Bright Publisher in Journal of Applied Data Sciences
- Vol. 6 (1) , 788-799
Open AccessHigh-Accuracy Stroke Detection System Using a CBAM-ResNet18 Deep Learning Model on Brain CT Images
High-Accuracy Stroke Detection System Using a CBAM-ResNet18 Deep Learning Model on Brain CT Images
- Open Access
- 25 January 2025
- Published by Bright Publisher in Journal of Applied Data Sciences
- Vol. 6 (1) , 753-772
Traditional-Enhance-Mobile-Ubiquitous-Smart: Model Innovation in Higher Education Learning Style Classification Using Multidimensional and Machine Learning Methods
Traditional-Enhance-Mobile-Ubiquitous-Smart: Model Innovation in Higher Education Learning Style Classification Using Multidimensional and Machine Learning Methods
- 25 January 2025
- Published by Bright Publisher in Journal of Applied Data Sciences
- Vol. 6 (1) , 753-772
Open AccessTraditional-Enhance-Mobile-Ubiquitous-Smart: Model Innovation in Higher Education Learning Style Classification Using Multidimensional and Machine Learning Methods
Traditional-Enhance-Mobile-Ubiquitous-Smart: Model Innovation in Higher Education Learning Style Classification Using Multidimensional and Machine Learning Methods