Early marriage remains a pressing issue among adolescents in Lombok, Indonesia, influenced by cultural norms, educational barriers, and economic challenges. This study develops an emotion classification and reason identification framework for a virtual counseling chatbot to support prevention efforts. Five functional emotion categories ‘Enthusiastic’, ‘Gentle’, ‘Analytical’, ‘Inspirational’, and ‘Cautionary’ were defined to capture counseling tones. The system leverages IndoBERT with a two-phase fine-tuning strategy. Phase 1 used a balanced dataset of 2,000 samples and achieved a macro F1-score of 0.95, while Phase 2 refined the model using 10,000 imbalanced pseudo-labeled samples, yielding a macro F1-score of 0.88 and improved sensitivity to minority classes. In addition, a semantic similarity-based reason identification module was implemented to classify user inputs into Education, Economy, Religion, or Culture categories, enhancing context awareness beyond simple keyword matching. Performance evaluation employed accuracy, precision, recall, and F1-score, supported by confusion matrices and training plots for generalization analysis. A descriptive emotion-to-gesture mapping was also designed to link each emotion category with static body pose visualizations, providing a conceptual basis for future multimodal applications.

Early marriage among individuals of immature age continues to draw significant attention in Lombok. As of 2021, the prevalence rate stands at 16.59%, indicating that this social issue remains unresolved within the region's community dynamics. Limited access to counseling services particularly in rural areas poses a significant barrier to prevention efforts. This study introduces a virtual counseling chatbot designed to detect and correct Indonesian language text errors during user interactions. The system integrates IndoRoBERTa for error detection and Mistral-7B-Instruct to refine speech to text transcriptions. IndoRoBERTa was trained on synthetic datasets to classify user input as accurate or incorrect, while Mistral-7B-Instruct generates context aware corrections. Achieving an accuracy rate of 98.90%, IndoRoBERTa outperformed benchmark models such as BERT and RNN. The proposed chatbot offers an adaptive and accessible digital solution, especially for communities with limited access to conventional counseling services. This approach highlights the potential of AI-driven tools to support early intervention strategies and reduce the incidence of child marriage in underserved regions.

This study investigates the optimization of Gravity Forward Modeling (GFM) by implementing an OpenMP parallel computing approach, focusing on the synthetic model and topography of Bawean Island. The research addresses computational limitations in traditional GFM by utilizing parallel processing techniques to enhance efficiency and resolution. By modeling subterranean layers with rectangular prisms and using the Okabe equation for calculations, the study significantly improves computational speed and resource management. The practical application on Bawean Island includes a strategic distribution of observation stations to facilitate Bouguer Anomaly (BA) computations, providing a clearer understanding of subsurface density anomalies. Notably, the study demonstrates a 600% increase in efficiency for the synthetic case when OpenMP with 8 threads is applied to the utilized architecture. For the field model, computations spanning 400 square kilometers on Bawean Island, encompassing over 1.3 billion GFM calculations, are completed in just 143 seconds. Additionally, the gravity terrain values on Bawean Island exhibit a range of up to 65 mGal, which closely correlates with the high-resolution topographic map. These findings highlight the considerable advantages of parallel computing in enhancing the efficiency and feasibility of complex geophysical modeling tasks, offering substantial improvements for large-scale geophysical exploration.

The use of email as a communication technology is now increasingly being exploited. Along with its progress, email spam problem becomes quite disturbing to email user. The resulting negative impacts make effective spam email detection techniques indispensable. A spam email detection algorithm or spam classifier will work effectively if supported by proper preprocessing steps (noise removal, stop words removal, stemming, lemmatization, term frequency). This research studies the effect of preprocessing steps on the performance of supervised spam classifier algorithms. Experiments were conducted on two widely used supervised spam classifier algorithms: Naïve Bayes and Support Vector Machine. The evaluation is performed on the Ling-spam corpus dataset and uses evaluation metrics: accuracy. The experimental results show that different preprocessing steps give different effects to different classifier.

<p>Penelitian ini memiliki tujuan untuk membuat sebuah model prediksi hasil produksi biogas pada reaktor tipe <em>batch</em>. Simulasi pencernaan anaerobik akan glukosa sebagai substrat utama dengan konsentrasi awal 500 mgCOD/l, dan simulasi akan dilakukan selama 120 jam. Dalam penelitian ini juga bertujuan untuk mengetahui konsentrasi mikroorganisme yang terlibat dalam proses pencernaan anaerobik, serta akan dilakukan beberapa analisis seperti perbandingan metana yang dihasilkan pada simulasi dan eksperimen, pengaruh jumlah iterasi terhadap waktu yang dibutuhkan untuk melakukan <em>running </em>program, perbandingan jumlah glukosa dan mikroorganisme yang digunakan dalam simulasi terhadap jumlah metana yang akan dihasilkan. Untuk memprediksi jumlah produksi biogas, terdapat sebuah model yang umum digunakan yaitu <em>Anaerobic Digestion Model</em> <em>No 1</em> (ADM1). ADM1 dikembangkan oleh Asosiasi Water International (IWA) pada tahun 2002. Agar mendapatkan model yang memiliki akurasi yang tinggi akan digunakan sebuah metode numerik yaitu Hamming <em>Predictor-Corrector</em>. Setelah simulasi pencernaan anaerobik dilakukan, metana yang dihasilkan sebesar 417,48 MgCOD/l. Lalu mikroorganisme glukosa mengalami pertumbuhan yang maksimum jika dibandingkan dengan mikroorganisme lain yaitu sebesar 77 MgCOD/l. Konsentrasi awal substrat glukosa dan konsentrasi mikroorganisme yang digunakan pada proses simulasi sangat berpengaruh terhadap jumlah metana yang dihasilkan. Namun untuk konsentrasi awal mikroba yang lebih dari 30 MgCOD/l, cenderung menghasilkan metana yang konstan.</p>

This paper focuses on the numerical modelling and simulation of tsunami waves triggered by an underwater landslide. The equation of motion for water waves is represented by the Nonlinear Shallow Water Equations (NSWE). Meanwhile, the motion of underwater landslide is modeled by incorporating a term for bottom motion into the NSWE. The model is solved numerically by using a finite volume method with a momentum conservative staggered grid scheme that is proposed by Stelling & Duinmeijer 2003 [12]. Here, we modify the scheme for the implementation of bottom motion. The accuracy of the implementation for representing wave runup and rundown is shown by performing the runup of harmonic wave as proposed by Carrier & Greenspan 1958 [2], and also solitary wave runup of Synolakis, 1986 [14], for both breaking and non-breaking cases. For the underwater landslide, result of the simulation is compared with simulation using the Boundary Integral Equation Model (BIEM) that is performed by Lynett and Liu, 2002 [9].

<br /><table class="data" width="100%"><tbody><tr valign="top"><td class="value">Angklung adalah kebudayaan khas Indonesia yang telah diakui UNESCO sejak tahun 2010. Salah satu bentuk pelestarian terhadap kebudayaan angklung, pemerintah kota Bandung menjadikan Saung Angklung Udjo sebagai tempat untuk melestarikan angklung. Seiring perkembangan teknologi, <em>game </em>angklung untuk <em>smartphone </em>sudah banyak dikembangkan. Pada suatu <em>game,</em> aspek <em>usability </em>merupakan salah satu aspek penting penghubung antara <em>user </em>dengan <em>game. </em>Berdasarkan evaluasi <em>usability </em>yang dilakukan menggunakan metode <em>Quality in Use Integrated Measurement </em>(QUIM) kepada beberapa aplikasi dan <em>game</em> angklung diperoleh rata- rata pengujian untuk semua faktor belum memenuhi aspek <em>usability </em>yang mengakibatkan pengguna mengalami kendala dalam menggunakan aplikasi <em>game </em>angklung. Oleh sebab itu perlu dilakukan perancangan <em>user interface </em>terhadap aplikasi <em>game</em> angklung supaya dapat digunakan dalam bermain angklung. Metode yang digunakan untuk perancangan <em>user interface </em>aplikasi <em>game </em>angklung adalah metode <em>user centered design </em>yang berfokus pada <em>characteristics, task, environments </em>dari pengguna. Berdasarkan hasil evaluasi dengan menggunakan QUIM aplikasi <em>game</em> Belajar Bermain Angklung dapat memenuhi kebutuhan dan kemampuan pengguna, hal ini terlihat dari nilai <em>usability</em> terendah sebesar 86% yang artinya sudah memenuhi aspek <em>usability</em></td></tr></tbody></table>

Klasifikasi gender berdasarkan suara mempengaruhi tingkat performansi dari identifikasi suara. oleh karena itu penelitian tugas akhir ini penulis membuat sistem klasifikasi gender berdasarkan suara. penelitian ini menggunakan dataset berupa rekaman suara sebanyak 2260 suara yang terbagi menjadi dua kelas yaitu "laki-laki" dan "perempuan". Untuk proses ekstrasi ciri digunakan Mel-Frequency Cepstral (MFCC) untuk menghasilkan vektor ciri dari sinyal suara. dalam proses klasifikasi digunakan metode Support Vector Machine (SVM). Hasil akurasi SVM tertinggi untuk klasifikasi gender berdasarkan suara dapat mencapai 100\% menggunakan kernel Polynomial dengan degree = 1.

Aerial images has different data characteristics when compared to other types of images. An aerial image usually contains small insignificant objects that can cause errors in the unsupervised segmentation method. K-means clustering, one of the widely used unsupervised image segmentation methods, is highly vulnerable to local optima. In this study, Adaptive Fireworks Algorithm (AFWA) is proposed as an alternative to the K-means algorithm in optimizing the clustering process in the cluster-based segmentation method. AFWA is then applied to perform aerial image segmentation and the results are compared with K-means. Based on the comparison using Probabilistic Rand Index (PRI) and Variation of Information (VI) evaluation metrics, AFWA produces an overall better segmentation quality.

<p><strong>Abstract</strong></p><p>Today, Software Defined Network (SDN) has been globally recognized as a new technology for network architecture. But, there is still lack in security. Many studies use methods such as the Intrusion Prevention System (IPS) and Intrusion Detection System (IDS) to deal with social problems. But there is still a lack of security in terms of network performance. To solve the problem, can be used Deep Packet Inspection method (DPII) which make administrators can directly know what happens to traffic in real time. In this research, DPI will be implemented as security method and tested with Denial of Service (DoS) attack with Direct Attack. The results of testing on SDN networks that have been added DPI can perform packet detection such as IDS and blocking such as IPS with good performance time in overcoming attack.</p><p><strong>Keywords: </strong>SDN, DPI, DoS attack, Direct Attack, performance</p>