Abstract:YouTube has become a major platform for public discourse in Indonesia, yet large-scale sentiment analysis of its comments remains challenging due to dynamic content, informal language, and limited labeled data. This study…
y proposes a Selenium–IndoBERT pipeline for sentiment analysis of Indonesian YouTube comments using a pseudo-labeling approach. Data were collected from ten YouTube videos discussing the One Piece flag phenomenon, yielding 10,842 comments after preprocessing. Selenium was employed to extract comments from dynamic pages, while IndoBERT was fine-tuned on a small manually labeled dataset and used to generate pseudo-labels for unlabeled data. Model performance was evaluated using probabilistic metrics, including Coverage, Expected Calibration Error (ECE), and Brier Score. At a confidence threshold of 0.75, 78.5% of comments received pseudo-labels, with an ECE of 0.095 and a Brier Score of 0.174. Manual validation showed substantial agreement with human annotations (Fleiss’ kappa = 0.72). The results indicate that the proposed pipeline enables scalable and reliable sentiment analysis with minimal manual annotation.
Abstract:Abstract: This study discusses the optimization of RAG for a FAQ system in the field of information technology product security certification at BSSN. Although LLM generate reliable responses, they often lack up-to-date…
and domain-specific knowledge, which can be addressed through the RAG approach. This research aims to optimize a domain-specific RAG system by improving embedding performance, enhancing prompt robustness, and increasing retrieval accuracy. The research methods consist of three stages. The first stage involves fine-tuning the bge-m3 embedding model and evaluating its performance using MRR, Recall, and AUC. The second stage applies prompt engineering techniques, namely the SRSM and Autodefense, to mitigate direct-injection and escape-character prompt injection attacks. The third stage evaluates the proposed RAG system using Precision, Recall, and F1-Score metrics against four baseline models. The results of research show that the fine-tuned embedding model achieves higher performance than the original model, with MRR@1 and Recall@1 values of 0.80 and an AUC@100 of 0.7023. In addition, the proposed prompt engineering techniques demonstrate robustness against prompt injection attacks, while the overall RAG system attains a perfect Precision, Recall, and F1-Score of 1.00. In conclusion, the proposed approach effectively enhances retrieval accuracy, embedding quality, and system security, resulting in a more reliable RAG-based FAQ system for information technology product security certification.
Keywords: embedding fine-tuning; large language model; prompt engineering; prompt injection mitigation; retrieval-augmented generation
Abstrak: Studi ini membahas optimasi RAG untuk sistem FAQ di bidang sertifikasi keamanan produk teknologi informasi di BSSN. Meskipun LLM menghasilkan respons yang andal, mereka seringkali kurang memiliki pengetahuan terkini dan spesifik domain, yang dapat diatasi melalui pendekatan RAG. Penelitian ini bertujuan untuk mengoptimalkan sistem RAG spesifik domain dengan meningkatkan kinerja embedding, meningkatkan ketahanan prompt dan meningkatkan akurasi pengambilan. Metode penelitian terdiri dari tiga tahap. Tahap pertama melibatkan fine-tuning model embedding bge-m3 dan mengevaluasi kinerjanya menggunakan Mean Reciprocal Rank (MRR), Recall, dan AUC. Tahap kedua menerapkan teknik rekayasa prompt, yaitu Self- SRSM dan Autodefense, untuk mengurangi serangan direct-injection dan escape-character prompt injection. Tahap ketiga mengevaluasi sistem RAG yang diusulkan menggunakan metrik Presisi, Recall, dan F1-Score terhadap empat model dasar. Hasil penelitian menunjukkan bahwa model embedding yang disempurnakan mencapai kinerja yang lebih tinggi daripada model asli, dengan nilai MRR@1 dan Recall@1 sebesar 0,80 dan AUC@100 sebesar 0,7023. Selain itu, teknik rekayasa prompt yang diusulkan menunjukkan ketahanan terhadap serangan injeksi prompt, sementara sistem RAG secara keseluruhan mencapai Presisi, Recall, dan F1-Score sempurna sebesar 1,00. Kesimpulannya, pendekatan yang diusulkan secara efektif meningkatkan akurasi pengambilan, kualitas embedding dan keamanan sistem, menghasilkan sistem FAQ berbasis RAG yang lebih andal untuk sertifikasi keamanan produk teknologi informasi.
Kata kunci: penyempurnaan embedding; model bahasa besar; rekayasa prompt; mitigasi injeksi prompt; retrieval-augmented generation
Abstract:Abstract: One of the main challenges in digital image processing is limited resolution, which makes it difficult to preserve visual details when images are enlarged. Conventional methods such as Bilinear Interpolation are…
e commonly used for image upscaling; however, these approaches often produce blurred images, lose fine textures, and fail to reconstruct complex visual structures. This study aims to enhance digital image resolution by employing a deep learni based approach using a Low-Light Convolutional Neural Network (LLCNN) built upon a Deep Neural Network (DNN) architecture. The dataset used in this study is the DIV2K dataset, which consists of 1,000 high-resolution images. These images were downsampled using scaling factors of ×2, ×3, and ×4 to generate paired Low Resolution–High Resolution (LR–HR) data for training and evaluation. The proposed LLCNN is designed to extract important features such as edges, textures, and local patterns through multiple convolutional layers, followed by non-linear mapping to reconstruct high-resolution images more accurately. Quantitative performance evaluation was conducted using the Peak Signal-to-Noise Ratio (PSNR) and the Structural Similarity Index (SSIM). Model performance was evaluated quantitatively using the Peak Signal-to-Noise Ratio (PSNR) metric. Experimental results showed that the proposed method improved image quality compared to the bilinear method. These results indicate that the deep learning based approach effectively improves image sharpness and structural fidelity, thereby demonstrating its potential for digital image resolution enhancement.
Keywords: deep neural network; image resolution; low-light convolutional neural network; machine learning
Abstrak: Permasalahan utama dalam pengolahan citra digital adalah keterbatasan resolusi yang menyebabkan detail visual sulit dipertahankan ketika citra diperbesar. Metode konvensional seperti Bilinear Interpolation masih banyak digunakan, namun sering menghasilkan citra buram, kehilangan tekstur halus, serta tidak mampu merekonstruksi struktur visual yang kompleks. Penelitian ini bertujuan untuk meningkatkan kualitas resolusi citra digital dengan memanfaatkan pendekatan deep learning berbasis Low-Light Convolutional Neural Network (LLCNN) yang dibangun di atas arsitektur Deep Neural Network (DNN). Data yang digunakan dalam penelitian ini berasal dari dataset DIV2K, yang terdiri dari 1000 citra beresolusi tinggi. Citra tersebut diturunkan menjadi resolusi rendah menggunakan faktor downsampling ×2, ×3, dan ×4 untuk membentuk pasangan data Low Resolution–High Resolution (LR–HR) sebagai data pelatihan dan pengujian. LLCNN dirancang untuk mengekstraksi fitur-fitur penting seperti tepi, tekstur, dan pola lokal melalui beberapa lapisan konvolusi, kemudian melakukan pemetaan non-linear guna merekonstruksi citra resolusi tinggi secara lebih presisi. Evaluasi performa model dilakukan secara kuantitatif menggunakan metrik Peak Signal-to-Noise Ratio (PSNR). Hasil eksperimen menunjukkan bahwa metode yang diusulkan mampu meningkatkan kualitas citra dibandingkan metode bilinear. Hasil ini membuktikan bahwa pendekatan berbasis deep learning efektif dalam meningkatkan ketajaman dan kesesuaian struktur citra digital.
Kata kunci: deep neural network; low-light convolutional neural network; machine learning; resolusi citra
Abstract:Abstract: Obesity is an escalating global health concern, with unhealthy lifestyle patterns contributing significantly to its development. This study aims to evaluate and compare three clustering techniques for categorizing…
ing lifestyle patterns and obesity-related factors: K-Means, Agglomerative Clustering, and Gaussian Mixture Model (GMM). The data used in this study is sourced from the Food Nutrition dataset, which includes variables such as dietary habits, physical activity, and socio-economic status. The three clustering methods were assessed using evaluation metrics such as Silhouette Score, Davies-Bouldin Index (DBI), and Calinski-Harabasz Index (CHI). The findings revealed that K-Means exhibited the best performance in terms of cluster separation with a Silhouette Score of 0.5559, while GMM showed better flexibility in handling more complex data. Although Agglomerative Clustering produced acceptable results, it had a higher overlap between clusters compared to the other methods. This study offers valuable insights into selecting the most appropriate clustering technique based on the data characteristics.
Keywords: agglomerative; clustering; GMM; k-means; lifestyle patterns; obesity
Abstrak: Obesitas menjadi masalah kesehatan yang semakin meningkat di seluruh dunia, dengan pola hidup yang tidak sehat berperan besar dalam perkembangannya. Penelitian ini bertujuan untuk membandingkan tiga metode clustering dalam mengelompokkan pola gaya hidup dan faktor yang memengaruhi obesitas, yaitu K-Means, Agglomerative Clustering, dan Gaussian Mixture Model (GMM). Data yang digunakan diperoleh dari dataset Food Nutrition yang mencakup informasi terkait pola makan, aktivitas fisik, serta faktor sosial-ekonomi. Ketiga metode tersebut diuji dengan menggunakan beberapa metrik evaluasi, seperti Silhouette Score, Davies-Bouldin Index (DBI), dan Calinski-Harabasz Index (CHI). Hasil penelitian menunjukkan bahwa K-Means memiliki kinerja terbaik dalam hal pemisahan klaster, dengan nilai Silhouette Score sebesar 0.5559, sementara GMM lebih fleksibel dalam menangani data yang lebih kompleks. Meskipun Agglomerative Clustering memberikan hasil yang dapat diterima, tumpang tindih antar klaster lebih besar dibandingkan dengan kedua metode lainnya. Penelitian ini memberikan pemahaman yang lebih baik mengenai pemilihan metode clustering yang tepat berdasarkan karakteristik data yang digunakan.
Kata kunci: agglomerative; clustering; GMM; k-means; obesitas; pola gaya hidup
Abstract:Abstract: Manual management of domain validity periods and SSL certificates is prone to human error and can cause service disruptions, as was the case at PT XYZ. A reactive approach that relies on vendor notifications has…
s proven to be insufficient to ensure operational continuity. This research aims to design and implement an automated monitoring system to transform this manual approach into a preventive and proactive security framework. The method used is the implementation of an open-source stack consisting of Prometheus to collect metrics from specialized exporters (Blackbox and Domain Exporter), and Grafana for informative centralized dashboard visualization. The system is also integrated with early warning notifications via Telegram for rapid incident response. The result is a functional system with a centralized dashboard that visually displays the remaining validity period of assets using color markers (green for safe status, yellow for early warning, and red for critical status). System testing showed very high accuracy, reaching 100% for domains (MAE 0 days) and 99.45% for SSL certificates (MAE 1.0 days). This system has successfully transformed manual processes into automated and preventive ones, significantly mitigating the risk of human error and ensuring the reliability of digital services.
Keywords: domain; grafana; monitoring; prometheus; SSL certificate.
Abstrak: Pengelolaan manual masa berlaku domain dan sertifikat SSL rentan terhadap human error dan dapat menyebabkan gangguan layanan, seperti yang pernah terjadi di PT XYZ. Pendekatan reaktif yang mengandalkan notifikasi vendor terbukti tidak lagi memadai untuk menjamin kontinuitas operasional. Penelitian ini bertujuan merancang dan mengimplementasikan sistem pemantauan otomatis untuk mentransformasi pendekatan manual tersebut menjadi kerangka kerja keamanan yang preventif dan proaktif. Metode yang digunakan adalah implementasi stack open-source yang terdiri dari Prometheus untuk mengumpulkan metrik dari exporter spesialis (Blackbox dan Domain Exporter), serta Grafana untuk visualisasi dasbor terpusat yang informatif. Sistem ini juga diintegrasikan dengan notifikasi peringatan dini melalui Telegram untuk respons insiden yang cepat. Hasilnya adalah sebuah sistem fungsional dengan dashboard terpusat yang menampilkan sisa masa berlaku aset secara visual menggunakan penanda warna (hijau untuk status aman, kuning untuk peringatan dini, dan merah untuk status kritis). Pengujian sistem menunjukkan akurasi yang sangat tinggi, mencapai 100% untuk domain (MAE 0 hari) dan 99.45% untuk sertifikat SSL (MAE 1.0 hari). Sistem ini berhasil mengubah proses manual menjadi otomatis dan preventif, secara signifikan memitigasi risiko human error dan menjamin keandalan layanan digital.
Kata kunci: domain; grafana; pemantauan; prometheus; sertifikat SSL.
Abstract:Abstract: Asthma is a chronic respiratory disease that affects millions of people worldwide, making early detection crucial to prevent complications. This study aims to compare the performance of the Decision Tree and Random…
ndom Forest algorithms in classifying asthma based on clinical symptom data. The data were processed through feature selection and model training stages, then evaluated using accuracy, precision, recall, and F1-score.The experimental analysis revealed that the Random Forest algorithm surpassed the Decision Tree in all metrics, achieving 95.19% accuracy, 90.43% precision, 95.00% recall, and 93.00% F1-score. In contrast, the Decision Tree obtained 89.14% accuracy, 90.60% precision, 88.70% recall, and 89.70% F1-score. These results suggest that Random Forest is more robust and dependable, especially in managing complex and imbalanced medical datasets.
Keywords: asthma detection; decision tree; random forest; machine learning.
Abstrak: Asma merupakan penyakit pernapasan kronis yang memengaruhi jutaan orang di seluruh dunia sehingga deteksi dini sangat penting untuk mencegah komplikasi. Penelitian ini bertujuan membandingkan kinerja algoritma Decision Tree dan Random Forest dalam mengklasifikasikan asma berdasarkan data gejala klinis. Data diproses melalui tahapan seleksi fitur dan pelatihan model, kemudian dievaluasi menggunakan akurasi, presisi, recall, dan F1-score. Hasil penelitian menunjukkan bahwa Random Forest memberikan performa terbaik dengan akurasi 90.43%, presisi 95.00%, recall 95.00%, dan F1-score 93.00%. Sebaliknya, Decision Tree memperoleh akurasi 89.14%, presisi 90.60%, recall 88.70%, dan F1-score 89.70%. Hasil ini menunjukkan bahwa Random Forest lebih kuat dan dapat diandalkan, terutama dalam mengelola kumpulan data medis yang kompleks dan tidak seimbang.
Kata kunci: deteksi asma; decision tree; random forest; pembelajaran mesin.
Abstract:Abstract: In the era of sensitive health data and frequent cyberattacks, securing electronic medical records (EMR) has become a critical challenge. This study proposes a hybrid encryption framework combining Affine and AES…
ES algorithms with an AI-based key management module to enhance EMR security while maintaining efficiency. A dataset of 1,000 simulated records was evaluated using five cryptographic configurations: Affine-only, AES-only, RSA-only, Affine–AES, and Affine–AES with AI. Performance was measured through encryption/decryption latency and ciphertext size, while security was assessed under brute-force, SQL injection, and phishing simulations. The AI decision tree for key generation was evaluated using accuracy, precision, recall, F1-score, and entropy metrics. Results show that the AI-enhanced hybrid method eliminates brute-force success, introduces only minor latency overhead, and generates high-entropy keys with reliability above 98%. These findings indicate that integrating AI-based dynamic key regeneration into hybrid encryption can improve EMR security while remaining practical for clinical and cloud-based healthcare systems. Future work should involve real clinical datasets and explore post-quantum cryptographic extensions.
Keywords: AI key management; attack resistance; encryption performance; electronic medical records; hybrid encryption
Abstrak: Di era meningkatnya sensitivitas data kesehatan dan maraknya serangan siber, perlindungan Rekam Medis Elektronik (RME) menjadi tantangan penting. Penelitian ini mengusulkan kerangka enkripsi hibrida yang menggabungkan algoritma Affine dan AES dengan modul manajemen kunci berbasis AI untuk meningkatkan keamanan RME tanpa mengorbankan efisiensi. Dataset simulasi berisi 1.000 entri diuji menggunakan lima konfigurasi kriptografi: Affine-only, AES-only, RSA-only, Affine–AES, serta Affine–AES dengan AI. Performa diukur melalui latensi enkripsi/dekripsi dan ukuran ciphertext, sedangkan keamanan dievaluasi melalui simulasi serangan brute force, SQL injection, dan phishing. Model decision tree untuk manajemen kunci dinilai menggunakan metrik akurasi, presisi, recall, F1-score, dan entropi. Hasil menunjukkan bahwa metode hibrida dengan AI menghilangkan keberhasilan brute force, menambah overhead latensi yang minimal, serta menghasilkan kunci berentropi tinggi dengan reliabilitas di atas 98%. Temuan ini menunjukkan bahwa regenerasi kunci dinamis berbasis AI dalam skema enkripsi hibrida dapat meningkatkan keamanan RME sekaligus tetap praktis untuk sistem klinis dan layanan kesehatan berbasis cloud. Penelitian selanjutnya disarankan menggunakan dataset klinis nyata dan mengeksplorasi kriptografi pascakuantum.
Kata kunci: enkripsi hibrida; ketahanan serangan; kinerja enkripsi; manajemen kunci berbasis AI; rekam medis elektronik
Abstract:Abstract: Recommendation systems are becoming increasingly important with the growth of streaming platforms. The purpose of this study is to compare the performance of Content-Based Filtering, Neural Collaborative Filtering,…
ing, and a combination of both in a movie recommendation system. The method used in this study involves retrieving movie details from the TMDB API and ratings from the MovieLens 32M Dataset (2010-2023). Each model's performance is evaluated using evaluation metrics such as RMSE and MAE. The results of this study indicate that Neural Collaborative Filtering achieves the best prediction performance (RMSE = 0.785423, MAE = 0.581262), followed by the hybrid model (RMSE = 0.800863, MAE = 0.660872), while Content-Based Filtering produces low performance and limits the capabilities of the hybrid model. In conclusion, these findings highlight the superiority of latent feature-based models such as NCF that learn directly from user interaction patterns over content-based approaches in the context of modern recommendation systems.
Keywords: content-based filtering; hybrid filtering; movie recommendation; neural collaborative filtering.
Abstrak: Sistem rekomendasi menjadi semakin penting seiring berkembangnya platform streaming. Tujuan dari penelitian ini adalah membandingkan kinerja Content-Based Filtering, Neural Collaborative Filtering dan kombinasi keduanya dalam sistem rekomendasi film. Metode yang digunakan dalam penelitian ini melibatkan pengambilan detail film dari TMDB API dan rating dari dataset MovieLens 32M Dataset (2010-2023). Setiap peforma model dievaluasi dengan menggunakan metrik evaluasi seperti RMSE dan MAE. Hasil dari penelitian ini menunjukkan bahwa Neural Collaborative Filtering mencapai kinerja prediksi terbaik (RMSE = 0.785423, MAE = 0.581262), diikuti oleh model hybrid (RMSE = 0.800863, MAE = 0.660872), sementara Content-Based Filtering menghasilkankan peforma yang rendah dan membatasi kemampuan model hybrid. Kesimpulannya, penelitian ini menyoroti superiotas model berbasis latent feature seperti NCF yang belajar langsung dari pola interaksi pengguna dibandingkan pendekatan berbasis konten dalam konteks sistem rekomendasi modern.
Kata kunci: content-based filtering; hybrid filtering; neural collaborative filtering; rekomendasi film.
Abstract:Abstract: Non-performing loans remain one of the main challenges faced by cooperatives, particularly when the loan eligibility assessment process is still conducted manually. This traditional approach tends to be time consuming,…
nsuming, subjective, and prone to inaccurate decisions. This study aims to develop a predictive model for borrower eligibility using the Support Vector Machine (SVM) algorithm as a more efficient and objective machine learning-based solution. A total of 1,000 loan history records were processed using RapidMiner software, taking into account variables such as salary, years of employment, loan amount, monthly installment, employment status, monthly expenses, number of dependents, housing status, age, and collateral value. The model’s performance was evaluated using a confusion matrix and classification metrics including accuracy, precision, recall, and kappa. The results indicate that the SVM model achieved an accuracy of 90.05%, precision of 90.13%, recall of 90.05%, and f1 score of 90,08%, reflecting a strong performance in classifying borrower eligibility. The application of this method makes a significant contribution to the development of data driven decision support systems within cooperative environments. This finding expands the scientific understanding in the field of microfinance and supports the implementation of artificial intelligence technologies in making decisions that are more precise, rapid, and accurate.
Keywords: cooperative; eligibility prediction; machine learning; non-performing loan; SVM
Abstrak: Kredit macet merupakan salah satu permasalahan utama yang dihadapi koperasi, terutama ketika proses penilaian kelayakan peminjam masih dilakukan secara manual. Pendekatan ini cenderung lambat, subjektif, dan berisiko menghasilkan keputusan yang kurang akurat. Penelitian ini bertujuan untuk membangun model prediksi kelayakan peminjam menggunakan algoritma Support Vector Machine (SVM) sebagai solusi berbasis machine learning yang lebih efisien dan objektif. Sebanyak 1.000 data riwayat pinjaman diolah menggunakan tools RapidMiner dengan mempertimbangkan variabel: gaji, lama bekerja, besar pinjaman, angsuran per bulan, status pegawai, pengeluaran bulanan, jumlah tanggungan, status rumah, umur, dan nilai jaminan. Evaluasi model dilakukan menggunakan confusion matrix dan metrik klasifikasi seperti akurasi, presisi, recall, dan kappa. Hasil menunjukkan bahwa model SVM mencapai akurasi 90,05%, presisi 90,13%, recall 90,05%, dan f1 score 90,08%, yang mencerminkan performa model yang sangat baik dalam mengklasifikasikan kelayakan peminjam. Penerapan metode ini memberikan kontribusi penting dalam pengembangan sistem pendukung keputusan berbasis data di lingkungan koperasi. Temuan ini memperluas wawasan keilmuan di bidang keuangan mikro dan mendukung penerapan teknologi kecerdasan buatan dalam pengambilan keputusan yang lebih tepat, cepat, dan akurat.
Kata Kunci: koperasi; kredit macet; machine learning; prediksi kelayakan; SVM
Abstract:Abstract: Face recognition based on deep learning has become an important technology in many areas. However, these systems often face challenges in real-world conditions, such as when the face is partially covered by accessories…
essories such as masks or glasses. This study aims to evaluate the effect of data augmentation by adding facial accessories (masks, glasses, and a combination of both) and geometric augmentation on the accuracy of face recognition systems. There are three types of datasets used in this method: the original dataset (category 1), the dataset with facial accessories augmentation (category 2), and the dataset with geometric augmentation (category 3). Data augmentation was performed on the training dataset to increase diversity, followed by the face detection process using SCRFD and feature extraction with ArcFace. The model was then trained using Multi-Layer Perceptron (MLP). Based on the results, adding face accessories (category 2) made the model a lot more accurate, hitting 99% accuracy. In category 3, adding geometric features improved accuracy to 91%. Other evaluation metrics, such as precision, recall, and F1-score, also showed improvement after augmentation. This study concludes that facial accessories augmentation is more effective in improving the accuracy and robustness of face recognition models compared to geometric augmentation.
Keywords: augmentation; deep learning; face recognition; glasses.
Abstrak: Pengenalan wajah berbasis deep learning telah menjadi salah satu teknologi penting dalam berbagai aplikasi. Namun, sistem ini sering kali menghadapi tantangan dalam kondisi dunia nyata, seperti saat wajah tertutup sebagian oleh aksesori seperti masker atau kacamata. Penelitian ini bertujuan untuk mengevaluasi pengaruh augmentasi data dengan menambahkan aksesori wajah (masker, kacamata, dan kombinasi keduanya) serta augmentasi geometris terhadap akurasi sistem pengenalan wajah. Metode yang digunakan melibatkan tiga kategori dataset: dataset asli tanpa augmentasi (kategori 1), dataset dengan augmentasi aksesoris wajah (kategori 2), dan dataset dengan augmentasi geometris (kategori 3). Augmentasi data dilakukan pada dataset pelatihan untuk meningkatkan keberagaman, diikuti dengan proses deteksi wajah menggunakan SCRFD dan ekstraksi fitur dengan ArcFace. Model kemudian dilatih menggunakan Multi-Layer Perceptron (MLP). Hasil penelitian menunjukkan bahwa augmentasi aksesoris wajah (kategori 2) memberikan peningkatan signifikan pada akurasi model, mencapai 99%, sedangkan kategori 3 dengan augmentasi geometris mencapai akurasi 91%. Metrik evaluasi lainnya, seperti precision, recall, dan F1-score, juga menunjukkan peningkatan setelah augmentasi. Penelitian ini menyimpulkan bahwa augmentasi aksesoris wajah lebih efektif dalam meningkatkan akurasi dan ketahanan model pengenalan wajah dibandingkan dengan augmentasi geometris.
Kata kunci: augmentasi; deep learning; kacamata; pengenalan wajah.