Dolly Indra
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Dolly Indra has a total of 74 co-authors. Here are the top co-authors by number of shared publications:
- Lilis Nur Hayati
Universitas Muslim Indonesia
4 shared publications Erick Irawadi Alwi
Universitas Muslim Indonesia
3 shared publications- Ramdan Satra
Universitas Muslim Indonesia
3 shared publications Muhammad Arfah Asis
2 shared publications- Tasrif Hasanuddin
Universitas Muslim Indonesia
2 shared publications - Abdul Rachman Manga
Universitas Muslim Indonesia
1 shared publication
Latest Publications
- Open Access
- 31 July 2026
- Published by Yocto Brain in Indonesian Journal of Data and Science
- Vol. 7 (2) , 238-246
Website-Based Boycott Product Detection System using Convolutional Neural Network: A comparison of YOLOv8 and VGG16
Introduction: Product boycotts have become a common form of social response in the digital era, yet manual identification of boycotted products can be slow and inaccurate. This study develops a website-based system and compares YOLOv8 and VGG16 for automated identification of boycott and non-boycott products from images. Method: A dataset of 4,250 food and beverage product images, comprising 3,638 boycott and 612 non-boycott products, was collected from internet sources and divided into 70% training, 20% validation, and 10% testing sets. YOLOv8 was trained using 640×640-pixel inputs as a one-stage object detector, while VGG16 used 224×224-pixel inputs with transfer learning as an image classifier. Both models were integrated into a website-based detection system and evaluated using accuracy, precision, recall, and F1-score. Results and Discussion: On 425 test images, YOLOv8 achieved 91.7% accuracy, 99.7% precision, 90.3% recall, and a 94.9% F1-score, substantially outperforming VGG16, which achieved 51.3% accuracy, 81.6% precision, 55.7% recall, and a 66.2% F1-score. YOLOv8 demonstrated greater robustness to variations in background, lighting, and product appearance because of its object-localization capability. Conclusion: YOLOv8 is more effective than VGG16 for website-based boycott product detection and provides a stronger foundation for practical real-time identification systems.
- 31 July 2026
- Published by Yocto Brain in Indonesian Journal of Data and Science
- Vol. 7 (2) , 238-246
Open AccessWebsite-Based Boycott Product Detection System using Convolutional Neural Network: A comparison of YOLOv8 and VGG16
Introduction: Product boycotts have become a common form of social response in the digital era, yet manual identification of boycotted products can be slow and inaccurate. This study develops a website-based system and compares YOLOv8 and VGG16 for automated identification of boycott and non-boycott products from images. Method: A dataset of 4,250 food and beverage product images, comprising 3,638 boycott and 612 non-boycott products, was collected from internet sources and divided into 70% training, 20% validation, and 10% testing sets. YOLOv8 was trained using 640×640-pixel inputs as a one-stage object detector, while VGG16 used 224×224-pixel inputs with transfer learning as an image classifier. Both models were integrated into a website-based detection system and evaluated using accuracy, precision, recall, and F1-score. Results and Discussion: On 425 test images, YOLOv8 achieved 91.7% accuracy, 99.7% precision, 90.3% recall, and a 94.9% F1-score, substantially outperforming VGG16, which achieved 51.3% accuracy, 81.6% precision, 55.7% recall, and a 66.2% F1-score. YOLOv8 demonstrated greater robustness to variations in background, lighting, and product appearance because of its object-localization capability. Conclusion: YOLOv8 is more effective than VGG16 for website-based boycott product detection and provides a stronger foundation for practical real-time identification systems.
- Open Access
- 29 June 2026
- Published by Institut Teknologi Nasional, Bandung in MIND Journal
- Vol. 11 (1) , 44-59
Klasifikasi Daun Herbal Berbasis Integrasi Fitur LBP dan Bentuk Menggunakan Random Forest
Abstrak Identifikasi daun herbal secara manual sering mengalami kendala akibat kemiripan visual antarspesies yang berpotensi menimbulkan kesalahan pemanfaatan. Penelitian ini bertujuan mengklasifikasikan sepuluh jenis daun herbal menggunakan kombinasi fitur tekstur Local Binary Pattern (LBP) dan fitur bentuk, yaitu aspect ratio, eccentricity, circularity, dan convexity. Dataset terdiri dari 500 citra yang dibagi menjadi 400 data latih dan 100 data uji dengan rasio 80:20. Tahap pre-processing meliputi resize, konversi ke grayscale, dan Gaussian Blur untuk mengurangi noise. Segmentasi dilakukan menggunakan Otsu thresholding untuk memperoleh objek daun dan Canny Edge Detection untuk menonjolkan struktur tekstur. Proses klasifikasi menerapkan algoritma Random Forest dengan pengujian beberapa kombinasi parameter guna memperoleh model optimal. Hasil terbaik diperoleh pada model dengan n_estimators=200, max_features=2, max_depth=none, dan min_samples_leaf=2, yang menghasilkan akurasi 92%, precision 92%, recall 92%, dan F1-score 92%. Kata kunci: Daun Tanaman Herbal; Fitur Bentuk; Fitur Tekstur; Klasifikasi Citra; Local Binary Pattern; Random Forest AbstractManual identification of herbal leaves often encounters challenges due to visual similarities among species, which can potentially lead to errors in their utilization. This study aims to classify ten types of herbal leaves using a combination of Local Binary Pattern (LBP) texture features and shape features, namely aspect ratio, eccentricity, circularity, and convexity. The dataset consists of 500 images divided into 400 training data and 100 testing data with an 80:20 ratio. The preprocessing stage includes resizing, grayscale conversion, and Gaussian Blur to reduce noise. Segmentation was performed using Otsu thresholding to extract the leaf object and Canny Edge Detection to enhance texture structures. The classification process applies the Random Forest algorithm with testing of several parameter combinations to obtain the optimal model. The best results were achieved with a model using n_estimators=200, max_features=2, max_depth=none, and min_samples_leaf=2, yielding an accuracy of 92%, precision of 92%, recall of 92%, and F1-score of 92%. Keywords: Herbal Plant Leaves, Texture Features, Shape Features, Image Classification, Local Binary Pattern; Random Forest
- 29 June 2026
- Published by Institut Teknologi Nasional, Bandung in MIND Journal
- Vol. 11 (1) , 44-59
Open AccessKlasifikasi Daun Herbal Berbasis Integrasi Fitur LBP dan Bentuk Menggunakan Random Forest
Abstrak Identifikasi daun herbal secara manual sering mengalami kendala akibat kemiripan visual antarspesies yang berpotensi menimbulkan kesalahan pemanfaatan. Penelitian ini bertujuan mengklasifikasikan sepuluh jenis daun herbal menggunakan kombinasi fitur tekstur Local Binary Pattern (LBP) dan fitur bentuk, yaitu aspect ratio, eccentricity, circularity, dan convexity. Dataset terdiri dari 500 citra yang dibagi menjadi 400 data latih dan 100 data uji dengan rasio 80:20. Tahap pre-processing meliputi resize, konversi ke grayscale, dan Gaussian Blur untuk mengurangi noise. Segmentasi dilakukan menggunakan Otsu thresholding untuk memperoleh objek daun dan Canny Edge Detection untuk menonjolkan struktur tekstur. Proses klasifikasi menerapkan algoritma Random Forest dengan pengujian beberapa kombinasi parameter guna memperoleh model optimal. Hasil terbaik diperoleh pada model dengan n_estimators=200, max_features=2, max_depth=none, dan min_samples_leaf=2, yang menghasilkan akurasi 92%, precision 92%, recall 92%, dan F1-score 92%. Kata kunci: Daun Tanaman Herbal; Fitur Bentuk; Fitur Tekstur; Klasifikasi Citra; Local Binary Pattern; Random Forest AbstractManual identification of herbal leaves often encounters challenges due to visual similarities among species, which can potentially lead to errors in their utilization. This study aims to classify ten types of herbal leaves using a combination of Local Binary Pattern (LBP) texture features and shape features, namely aspect ratio, eccentricity, circularity, and convexity. The dataset consists of 500 images divided into 400 training data and 100 testing data with an 80:20 ratio. The preprocessing stage includes resizing, grayscale conversion, and Gaussian Blur to reduce noise. Segmentation was performed using Otsu thresholding to extract the leaf object and Canny Edge Detection to enhance texture structures. The classification process applies the Random Forest algorithm with testing of several parameter combinations to obtain the optimal model. The best results were achieved with a model using n_estimators=200, max_features=2, max_depth=none, and min_samples_leaf=2, yielding an accuracy of 92%, precision of 92%, recall of 92%, and F1-score of 92%. Keywords: Herbal Plant Leaves, Texture Features, Shape Features, Image Classification, Local Binary Pattern; Random Forest
- Open Access
- 30 May 2026
- Published by LPPM ITN Malang in JATI (Jurnal Mahasiswa Teknik Informatika)
- Vol. 10 (3) , 5529-5536
ANALISIS SENTIMEN TERHADAP KANG DEDI MULYADI PADA MEDIA SOSIAL X MENGGUNAKAN ALGORITMA NAÏVE BAYES DAN SUPPORT VECTOR MACHINE (SVM)
Platform X menjadi ruang utama bagi masyarakat dalam mengekspresikan opini terhadap tokoh publik, termasuk Kang Dedi Mulyadi. Penelitian ini bertujuan menganalisis sentimen publik sekaligus menjawab keterbatasan studi sebelumnya yang umumnya berfokus pada isu politik nasional atau hanya menggunakan satu algoritma klasifikasi, sehingga perbandingan kinerja model pada konteks figur politik daerah masih terbatas. Data penelitian terdiri dari 7.190 tweet yang dikumpulkan melalui crawling selama Februari–September 2025. Data diproses melalui tahapan pre-processing, kemudian dilabeli otomatis menggunakan InSet Lexicon ke dalam sentimen positif dan negatif. Ekstraksi fitur dilakukan dengan TF-IDF menggunakan konfigurasi n-gram (1,2), dengan pembagian data 80:20 untuk pelatihan dan pengujian. Hasil pengujian menunjukkan bahwa SVM Linear memberikan performa terbaik dengan akurasi 0,8982, presisi 0,8991, recall 0,8970, dan F1-score 0,8977. Sementara itu, Naïve Bayes menghasilkan akurasi 0,8287, presisi 0,8297, recall 0,8268, dan F1-score 0,8276. Temuan ini menunjukkan bahwa SVM lebih efektif dibandingkan Naïve Bayes dalam mengklasifikasikan sentimen publik pada data media sosial X, serta memberikan kontribusi empiris pada kajian analisis sentimen tokoh politik berbasis media sosial
- 30 May 2026
- Published by LPPM ITN Malang in JATI (Jurnal Mahasiswa Teknik Informatika)
- Vol. 10 (3) , 5529-5536
Open AccessANALISIS SENTIMEN TERHADAP KANG DEDI MULYADI PADA MEDIA SOSIAL X MENGGUNAKAN ALGORITMA NAÏVE BAYES DAN SUPPORT VECTOR MACHINE (SVM)
Platform X menjadi ruang utama bagi masyarakat dalam mengekspresikan opini terhadap tokoh publik, termasuk Kang Dedi Mulyadi. Penelitian ini bertujuan menganalisis sentimen publik sekaligus menjawab keterbatasan studi sebelumnya yang umumnya berfokus pada isu politik nasional atau hanya menggunakan satu algoritma klasifikasi, sehingga perbandingan kinerja model pada konteks figur politik daerah masih terbatas. Data penelitian terdiri dari 7.190 tweet yang dikumpulkan melalui crawling selama Februari–September 2025. Data diproses melalui tahapan pre-processing, kemudian dilabeli otomatis menggunakan InSet Lexicon ke dalam sentimen positif dan negatif. Ekstraksi fitur dilakukan dengan TF-IDF menggunakan konfigurasi n-gram (1,2), dengan pembagian data 80:20 untuk pelatihan dan pengujian. Hasil pengujian menunjukkan bahwa SVM Linear memberikan performa terbaik dengan akurasi 0,8982, presisi 0,8991, recall 0,8970, dan F1-score 0,8977. Sementara itu, Naïve Bayes menghasilkan akurasi 0,8287, presisi 0,8297, recall 0,8268, dan F1-score 0,8276. Temuan ini menunjukkan bahwa SVM lebih efektif dibandingkan Naïve Bayes dalam mengklasifikasikan sentimen publik pada data media sosial X, serta memberikan kontribusi empiris pada kajian analisis sentimen tokoh politik berbasis media sosial
- Open Access
- 31 March 2026
- Published by Yocto Brain in Indonesian Journal of Data and Science
- Vol. 7 (1) , 105-118
Comparing Sentiment Labeling with RoBERTa and IndoBERTweet on Public Opinion of Program Makan Bergizi Gratis
The Program Makan Bergizi Gratis (MBG) is a flagship program of the Prabowo Subianto administration launched in 2024, triggering diverse public responses on social media. Sentiment analysis using deep learning models offers an effective approach to understanding public opinion at scale. However, selecting the appropriate model for Indonesian social media text remains challenging. This study aims to compare the performance of two pretrained transformer models, RoBERTa Base and IndoBERTweet Base, in conducting automatic sentiment labeling on Indonesian tweets related to the MBG program using a zero-shot labeling approach without human-annotated ground truth. A total of 1,831 tweets were collected from platform X and preprocessed using case folding, normalization, and stopword removal. Both models were applied in parallel to label each tweet with sentiment categories (positive, neutral, negative) along with confidence scores. The comparison was evaluated using agreement rate, Cohen's Kappa, and confidence score analysis. RoBERTa Base exhibits a conservative tendency with 75.20% neutral labels, while IndoBERTweet Base produces a more balanced distribution (68.16% neutral). The comparison shows 77.28% agreement with Cohen's Kappa of 0.490 (Moderate Agreement). RoBERTa Base achieves higher confidence (mean: 0.9559, 83.01% above 0.95) compared to IndoBERTweet Base (mean: 0.9236, 68.65% above 0.95). IndoBERTweet Base is more effective in detecting negative sentiment, identifying nearly twice as many negative tweets (13.54% vs. 7.48%). This study recommends IndoBERTweet Base for exploratory research requiring sensitive sentiment detection and RoBERTa Base for precision-critical applications. An ensemble approach combining both models is recommended for production-critical applications
- 31 March 2026
- Published by Yocto Brain in Indonesian Journal of Data and Science
- Vol. 7 (1) , 105-118
Open AccessComparing Sentiment Labeling with RoBERTa and IndoBERTweet on Public Opinion of Program Makan Bergizi Gratis
The Program Makan Bergizi Gratis (MBG) is a flagship program of the Prabowo Subianto administration launched in 2024, triggering diverse public responses on social media. Sentiment analysis using deep learning models offers an effective approach to understanding public opinion at scale. However, selecting the appropriate model for Indonesian social media text remains challenging. This study aims to compare the performance of two pretrained transformer models, RoBERTa Base and IndoBERTweet Base, in conducting automatic sentiment labeling on Indonesian tweets related to the MBG program using a zero-shot labeling approach without human-annotated ground truth. A total of 1,831 tweets were collected from platform X and preprocessed using case folding, normalization, and stopword removal. Both models were applied in parallel to label each tweet with sentiment categories (positive, neutral, negative) along with confidence scores. The comparison was evaluated using agreement rate, Cohen's Kappa, and confidence score analysis. RoBERTa Base exhibits a conservative tendency with 75.20% neutral labels, while IndoBERTweet Base produces a more balanced distribution (68.16% neutral). The comparison shows 77.28% agreement with Cohen's Kappa of 0.490 (Moderate Agreement). RoBERTa Base achieves higher confidence (mean: 0.9559, 83.01% above 0.95) compared to IndoBERTweet Base (mean: 0.9236, 68.65% above 0.95). IndoBERTweet Base is more effective in detecting negative sentiment, identifying nearly twice as many negative tweets (13.54% vs. 7.48%). This study recommends IndoBERTweet Base for exploratory research requiring sensitive sentiment detection and RoBERTa Base for precision-critical applications. An ensemble approach combining both models is recommended for production-critical applications
- Open Access
- 20 March 2026
- Published by LPPM ITN Malang in JATI (Jurnal Mahasiswa Teknik Informatika)
- Vol. 10 (2) , 3294-3301
ANALISIS SENTIMEN ULASAN APLIKASI BYOND BY BSI MENGGUNAKAN ALGORIMA NAÏVE BAYES DAN SUPPORT VECTOR MACHINE
Pemanfaatan aplikasi perbankan digital menjadi bagian penting dalam mendukung produktivitas masyarakat melalui layanan perbankan yang efisien dan praktis. Byond by BSI sebagai salah satu mobile banking turut berperan dalam memenuhi kebutuhan layanan keuangan nasabahnya. Seiring meningkatnya jumlah pengguna aplikasi Byond by BSI, muncul beragam opini yang belum terpetakan secara sistematis. Hal ini menciptakan kebutuhan akan analisis sistematis agar ulasan tersebut dapat dimanfaatkan sebagai basis evaluasi fungsionalitas aplikasi dan optimalisasi pengalaman pengguna. Penelitian ini bertujuan mengidentifikasi distribusi sentimen positif dan negatif pada ulasan pengguna aplikasi Byond by BSI serta menentukan algoritma terbaik berdasarkan nilai akurasi, presisi, recall, dan F1-score dalam mengklasifikasikan sentimen ulasan pengguna. Metode yang diterapkan untuk menentukan kategori sentimen ulasan adalah algoritma Naïve Bayes dan SVM. Hasil evaluasi penelitian menunjukkan bahwa algoritma SVM mampu menunjukkan kinerja lebih baik daripada Naïve Bayes dalam mengklasifikasikan ulasan. SVM menunjukkan kinerja terbaik pada pembagian data 90:10 dengan akurasi 95.33%, presisi 97.12%, recall 94.94%, dan f1-score 96%. sedangkan Naïve Bayes mencapai kinerja terbaiknya pada pembagian data 70:30 dengan akurasi 94%, presisi 95.10%, recall 94.45%, dan f1-score 94.81%.
- 20 March 2026
- Published by LPPM ITN Malang in JATI (Jurnal Mahasiswa Teknik Informatika)
- Vol. 10 (2) , 3294-3301
Open AccessANALISIS SENTIMEN ULASAN APLIKASI BYOND BY BSI MENGGUNAKAN ALGORIMA NAÏVE BAYES DAN SUPPORT VECTOR MACHINE
Pemanfaatan aplikasi perbankan digital menjadi bagian penting dalam mendukung produktivitas masyarakat melalui layanan perbankan yang efisien dan praktis. Byond by BSI sebagai salah satu mobile banking turut berperan dalam memenuhi kebutuhan layanan keuangan nasabahnya. Seiring meningkatnya jumlah pengguna aplikasi Byond by BSI, muncul beragam opini yang belum terpetakan secara sistematis. Hal ini menciptakan kebutuhan akan analisis sistematis agar ulasan tersebut dapat dimanfaatkan sebagai basis evaluasi fungsionalitas aplikasi dan optimalisasi pengalaman pengguna. Penelitian ini bertujuan mengidentifikasi distribusi sentimen positif dan negatif pada ulasan pengguna aplikasi Byond by BSI serta menentukan algoritma terbaik berdasarkan nilai akurasi, presisi, recall, dan F1-score dalam mengklasifikasikan sentimen ulasan pengguna. Metode yang diterapkan untuk menentukan kategori sentimen ulasan adalah algoritma Naïve Bayes dan SVM. Hasil evaluasi penelitian menunjukkan bahwa algoritma SVM mampu menunjukkan kinerja lebih baik daripada Naïve Bayes dalam mengklasifikasikan ulasan. SVM menunjukkan kinerja terbaik pada pembagian data 90:10 dengan akurasi 95.33%, presisi 97.12%, recall 94.94%, dan f1-score 96%. sedangkan Naïve Bayes mencapai kinerja terbaiknya pada pembagian data 70:30 dengan akurasi 94%, presisi 95.10%, recall 94.45%, dan f1-score 94.81%.
- Open Access
- 31 December 2025
- Published by Yocto Brain in Indonesian Journal of Data and Science
- Vol. 6 (3) , 378-392
Smart Waste Bin Prototype for University Waste Management
Background: Waste mismanagement remains a critical issue in Indonesian campuses, where ineffective segregation and collection practices contribute to environmental pollution. Smart technologies offer opportunities to improve waste handling efficiency and monitoring in university environments. Methods: This study developed a smart waste bin prototype that integrates Internet of Things (IoT) sensors, machine learning–based image classification (MobileNetV2 with TensorFlow Lite), GPS tracking, and LoRa communication. The system was designed to classify three types of waste—plastic bottles, snack packaging, and cans—while enabling fill-level monitoring, automated sorting, and real-time location reporting. Results: Experimental results showed strong classification accuracy for plastic bottles (100%), but lower performance for snack packaging (53–80%) and cans (40–67%), especially in low-light conditions or with darker materials. The overall real-time testing accuracy reached 45.1%. LoRa communication provided long-range connectivity but was affected by electromagnetic interference, while GPS tracking was reliable in open areas but inconsistent indoors. Conclusions: The prototype demonstrates the feasibility of integrating AI and IoT for scalable campus waste management. Despite environmental and hardware limitations, it offers a modular framework that can be refined with improved lighting, EMI shielding, and enhanced datasets. This research contributes a practical model for smart campus initiatives and supports the adoption of sustainable waste management practices in higher education environments.
- 31 December 2025
- Published by Yocto Brain in Indonesian Journal of Data and Science
- Vol. 6 (3) , 378-392
Open AccessSmart Waste Bin Prototype for University Waste Management
Background: Waste mismanagement remains a critical issue in Indonesian campuses, where ineffective segregation and collection practices contribute to environmental pollution. Smart technologies offer opportunities to improve waste handling efficiency and monitoring in university environments. Methods: This study developed a smart waste bin prototype that integrates Internet of Things (IoT) sensors, machine learning–based image classification (MobileNetV2 with TensorFlow Lite), GPS tracking, and LoRa communication. The system was designed to classify three types of waste—plastic bottles, snack packaging, and cans—while enabling fill-level monitoring, automated sorting, and real-time location reporting. Results: Experimental results showed strong classification accuracy for plastic bottles (100%), but lower performance for snack packaging (53–80%) and cans (40–67%), especially in low-light conditions or with darker materials. The overall real-time testing accuracy reached 45.1%. LoRa communication provided long-range connectivity but was affected by electromagnetic interference, while GPS tracking was reliable in open areas but inconsistent indoors. Conclusions: The prototype demonstrates the feasibility of integrating AI and IoT for scalable campus waste management. Despite environmental and hardware limitations, it offers a modular framework that can be refined with improved lighting, EMI shielding, and enhanced datasets. This research contributes a practical model for smart campus initiatives and supports the adoption of sustainable waste management practices in higher education environments.
- Open Access
- 27 October 2025
- Published by Indonesian Digital Technology University in Jurnal Pengabdian Masyarakat - Teknologi Digital Indonesia.
- Vol. 4 (2) , 158-165
Meningkatkan Daya Saing Pengrajin Anyaman Lontara Kana Tojeng Melalui Edukasi Penggunaan Media Sosial di Desa Bontokassi
Kegiatan pengabdian ini berhasil Meningkatkan Daya Saing Pengrajin Anyaman Lontara Kana Tojeng Melalui Edukasi Penggunaan Media Sosial di Desa Bontokassi di Desa Bontokassi. Berdasarkan hasil evaluasi menggunakan instrumen pre-test dan post-test yang terdiri dari 10 butir pertanyaan, dengan koresponden 20 anggota pengerajin Kana Tojeng terjadi peningkatan pemahaman yang signifikan. Skor rata-rata pre-test peserta berada pada angka 45,8%, sedangkan nilai post-test meningkat menjadi 75,5%, yang menunjukkan efektivitas pendekatan edukatif-partisipatif dalam mentransfer pengetahuan Penggunaan Media Sosial mitra. Memperlihatkan perbandingan yang jelas antara sebelum dan sesudah sosialisasi, yang turut memperkuat bukti pencapaian program. Di samping peningkatan kognitif, mitra juga menunjukkan perubahan Penggunaan Media Sosia untuk produk kerajinan. Secara umum, hasil kegiatan ini tidak hanya memperkuat daya saing produk lokal berbasis budaya, tetapi juga membuka peluang baru untuk ekspansi digital melalui e-commerce dan kolaborasi lebih luas di masa mendatang.
- 27 October 2025
- Published by Indonesian Digital Technology University in Jurnal Pengabdian Masyarakat - Teknologi Digital Indonesia.
- Vol. 4 (2) , 158-165
Open AccessMeningkatkan Daya Saing Pengrajin Anyaman Lontara Kana Tojeng Melalui Edukasi Penggunaan Media Sosial di Desa Bontokassi
Kegiatan pengabdian ini berhasil Meningkatkan Daya Saing Pengrajin Anyaman Lontara Kana Tojeng Melalui Edukasi Penggunaan Media Sosial di Desa Bontokassi di Desa Bontokassi. Berdasarkan hasil evaluasi menggunakan instrumen pre-test dan post-test yang terdiri dari 10 butir pertanyaan, dengan koresponden 20 anggota pengerajin Kana Tojeng terjadi peningkatan pemahaman yang signifikan. Skor rata-rata pre-test peserta berada pada angka 45,8%, sedangkan nilai post-test meningkat menjadi 75,5%, yang menunjukkan efektivitas pendekatan edukatif-partisipatif dalam mentransfer pengetahuan Penggunaan Media Sosial mitra. Memperlihatkan perbandingan yang jelas antara sebelum dan sesudah sosialisasi, yang turut memperkuat bukti pencapaian program. Di samping peningkatan kognitif, mitra juga menunjukkan perubahan Penggunaan Media Sosia untuk produk kerajinan. Secara umum, hasil kegiatan ini tidak hanya memperkuat daya saing produk lokal berbasis budaya, tetapi juga membuka peluang baru untuk ekspansi digital melalui e-commerce dan kolaborasi lebih luas di masa mendatang.
- Open Access
- 2 October 2025
- Published by Fakultas Sains dan Teknologi, Unira Malang in G-Tech: Jurnal Teknologi Terapan
- Vol. 9 (4) , 1878-1888
Sentiment Analysis of Public Opinion on Deforestation in Papua on YouTube Platform Using Long Short-Term Memory (LSTM) Method
Deforestation in Papua has emerged as a significant environmental concern, attracting considerable attention due to its effects on biodiversity and the livelihoods of indigenous communities. This study seeks to examine public sentiment toward the issue by analyzing comments posted on the YouTube platform, employing the Long Short-Term Memory (LSTM) method. A dataset of 3,000 comments was gathered and processed through several stages, including text cleaning, tokenization, normalization, and Term Frequency–Inverse Document Frequency (TF-IDF) weighting. Subsequently, an LSTM model was developed and assessed using accuracy, precision, recall, and F1-score as evaluation metrics. The results reveal that the LSTM model achieved an accuracy of 88.43%, a precision of 90.01%, a recall of 97.49%, and an F1-score of 93.60%. Nevertheless, signs of overfitting were observed, indicated by lower validation performance compared to training results. These findings demonstrate that the LSTM approach is effective for identifying public opinion regarding deforestation and can serve as a valuable reference in decision-making and the formulation of environmental policies.
- 2 October 2025
- Published by Fakultas Sains dan Teknologi, Unira Malang in G-Tech: Jurnal Teknologi Terapan
- Vol. 9 (4) , 1878-1888
Open AccessSentiment Analysis of Public Opinion on Deforestation in Papua on YouTube Platform Using Long Short-Term Memory (LSTM) Method
Deforestation in Papua has emerged as a significant environmental concern, attracting considerable attention due to its effects on biodiversity and the livelihoods of indigenous communities. This study seeks to examine public sentiment toward the issue by analyzing comments posted on the YouTube platform, employing the Long Short-Term Memory (LSTM) method. A dataset of 3,000 comments was gathered and processed through several stages, including text cleaning, tokenization, normalization, and Term Frequency–Inverse Document Frequency (TF-IDF) weighting. Subsequently, an LSTM model was developed and assessed using accuracy, precision, recall, and F1-score as evaluation metrics. The results reveal that the LSTM model achieved an accuracy of 88.43%, a precision of 90.01%, a recall of 97.49%, and an F1-score of 93.60%. Nevertheless, signs of overfitting were observed, indicated by lower validation performance compared to training results. These findings demonstrate that the LSTM approach is effective for identifying public opinion regarding deforestation and can serve as a valuable reference in decision-making and the formulation of environmental policies.
- Open Access
- 18 July 2025
- Published by Nak-Kanak: Journal of Child Research in Jurnal Ilmiah Kursor
- Vol. 13 (1) , 13-23
Generative adversarial networks (GANS) for generating face images
The advancement of artificial intelligence technology, particularly deep learning, presents significant potential in facial image processing. Generative Adversarial Networks (GANs), a type of deep learning model, have demonstrated remarkable capabilities in generating high-quality synthetic images through a competitive training process between a generator, which creates new data, and a discriminator, which evaluates its authenticity. However, the use of public facial datasets such as CelebA and FFHQ faces limitations in representing global demographic diversity and raises privacy concerns. This study aims to generate realistic synthetic facial datasets using the StyleGAN2-ADA architecture, a specialized variant of GAN, with two training approaches: training from scratch on two types of datasets (private and public), each containing 480 images. The public dataset used is FFHQ (Flickr-Faces-HQ), known for its broader facial variation and high-quality images. Evaluation is conducted using the Frechet Inception Distance (FID), a metric that assesses image quality by comparing the feature distributions of real and generated images. Results indicate that training from scratch with the public dataset (FFHQ) using a batch size of 16 and a learning rate of 0.0025 achieves an FID score of 85.67 and performance of 86.46% at Tick 100, whereas the private dataset, under the same conditions, results in an FID score of 98.59 with a performance of 18.54%.. The training from scratch approach with the public dataset proves more effective in generating high-quality synthetic facial images compared to the private dataset. In conclusion, this approach supports the optimal generation of realistic synthetic facial data.
- 18 July 2025
- Published by Nak-Kanak: Journal of Child Research in Jurnal Ilmiah Kursor
- Vol. 13 (1) , 13-23
Open AccessGenerative adversarial networks (GANS) for generating face images
The advancement of artificial intelligence technology, particularly deep learning, presents significant potential in facial image processing. Generative Adversarial Networks (GANs), a type of deep learning model, have demonstrated remarkable capabilities in generating high-quality synthetic images through a competitive training process between a generator, which creates new data, and a discriminator, which evaluates its authenticity. However, the use of public facial datasets such as CelebA and FFHQ faces limitations in representing global demographic diversity and raises privacy concerns. This study aims to generate realistic synthetic facial datasets using the StyleGAN2-ADA architecture, a specialized variant of GAN, with two training approaches: training from scratch on two types of datasets (private and public), each containing 480 images. The public dataset used is FFHQ (Flickr-Faces-HQ), known for its broader facial variation and high-quality images. Evaluation is conducted using the Frechet Inception Distance (FID), a metric that assesses image quality by comparing the feature distributions of real and generated images. Results indicate that training from scratch with the public dataset (FFHQ) using a batch size of 16 and a learning rate of 0.0025 achieves an FID score of 85.67 and performance of 86.46% at Tick 100, whereas the private dataset, under the same conditions, results in an FID score of 98.59 with a performance of 18.54%.. The training from scratch approach with the public dataset proves more effective in generating high-quality synthetic facial images compared to the private dataset. In conclusion, this approach supports the optimal generation of realistic synthetic facial data.
- Open Access
- 27 December 2024
- Published by P3M Politeknik Negeri Banjarmasin in Jurnal ELTIKOM
- Vol. 8 (2) , 91-99
Analysis of Hybrid Learning Sentiment among Information Systems Students using The Naïve Bayes Classifier
Hybrid learning, which combines online and face-to-face instruction, has gained significant attention. Particularly in the Faculty of Computer Science, student engagement in hybrid learning is a central concern that arises during implementation. Hybrid, or blended learning, integrates various teaching methods, such as face-to-face, computer-based, and mobile learning, and offers advantages by reducing the time required for meetings and information delivery. Sentiment analysis, a branch of text mining, aims to determine public opinion or sentiment on topics, events, or issues. This study surveyed 112 Information Systems students using an online questionnaire to assess their responses to hybrid learning, classified as positive, negative, or neutral using the Naïve Bayes classifier. The research stages included data collection, preprocessing, Naïve Bayes model training, model evaluation, and sentiment analysis. The study aimed to analyze hybrid learning’s impact on students' learning experiences and assess the accuracy of the Naïve Bayes method in classifying sentiments regarding this impact. The results indicated that the initial test had an accuracy of 60.87% without using the SMOTE up-sampling operator, while the second test achieved 80.65% accuracy with the operator.
- 27 December 2024
- Published by P3M Politeknik Negeri Banjarmasin in Jurnal ELTIKOM
- Vol. 8 (2) , 91-99
Open AccessAnalysis of Hybrid Learning Sentiment among Information Systems Students using The Naïve Bayes Classifier
Hybrid learning, which combines online and face-to-face instruction, has gained significant attention. Particularly in the Faculty of Computer Science, student engagement in hybrid learning is a central concern that arises during implementation. Hybrid, or blended learning, integrates various teaching methods, such as face-to-face, computer-based, and mobile learning, and offers advantages by reducing the time required for meetings and information delivery. Sentiment analysis, a branch of text mining, aims to determine public opinion or sentiment on topics, events, or issues. This study surveyed 112 Information Systems students using an online questionnaire to assess their responses to hybrid learning, classified as positive, negative, or neutral using the Naïve Bayes classifier. The research stages included data collection, preprocessing, Naïve Bayes model training, model evaluation, and sentiment analysis. The study aimed to analyze hybrid learning’s impact on students' learning experiences and assess the accuracy of the Naïve Bayes method in classifying sentiments regarding this impact. The results indicated that the initial test had an accuracy of 60.87% without using the SMOTE up-sampling operator, while the second test achieved 80.65% accuracy with the operator.