Claim Missing Document
Check
Articles

Found 18 Documents
Search

DETEKSI ANOMALI PADA AUDIT BARANG MILIK DAERAH MENGGUNAKAN EXTREME GRADIENT BOOSTING TEROPTIMASI BAYESIAN Sulman Edi S; Giat Karyono; Imam Tahyudin
Rabit : Jurnal Teknologi dan Sistem Informasi Univrab Vol 11 No 2 (2026): Juli
Publisher : LPPM Universitas Abdurrab

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36341/rabit.v11i2.8285

Abstract

The audit and reconciliation of Local Government-Owned Assets (BMD) constitute a critical governance process that remains heavily reliant on manual verification, rendering it susceptible to inefficiency and human error, particularly within datasets exhibiting extreme heterogeneity in acquisition values. This study proposes an intelligent analytical framework for asset depreciation anomaly detection using the Extreme Gradient Boosting (XGBoost) algorithm, optimized via a Bayesian Optimization approach through the Optuna framework. From an initial raw population of 98,526 records, the data underwent preprocessing to yield 42,241 clean records with unique profiles. To address the disparity in price ranges, the dataset was divided into three strata using the Equal Frequency quantile method, with the prediction target transformed into a depreciation ratio. The evaluation demonstrated highly precise performance, consistently achieving a coefficient of determination (R²) above 0.99. Bayesian optimization reduced the Weighted Average Percentage Error (WAPE) to a range of 0.54% to 1.32%. Using a 10% deviation threshold, the system automatically extracted 101 anomalous records (1.20%) from 8,450 test samples as red flags. The results confirm that this framework is highly viable as a decision-support instrument for public asset audits, in compliance with regional regulations.
3D word embedding vector feature extraction and hybrid CNN-LSTM for natural disaster reports identification Mohammad Reza Faisal; Dodon Turianto Nugrahadi; Irwan Budiman; Muliadi Muliadi; Mera Kartika Delimayanti; Septyan Eka Prastya; Imam Tahyudin
TELKOMNIKA (Telecommunication Computing Electronics and Control) Vol 22, No 5: October 2024
Publisher : Universitas Ahmad Dahlan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.12928/telkomnika.v22i5.26091

Abstract

Social media contain various information, such as natural disaster reports. Artificial intelligence is used to identify reports from eyewitnesses early for disaster warning systems. The artificial intelligence system includes a text classification model with feature extraction and classification algorithms. Word embedding-based feature extraction is widely used for 1-dimensional (1D) and 2-dimensional (2D) data, suitable for traditional or deep learning algorithms. However, applying feature extraction to 3-dimensional (3D) data for text classification is limited. Previous studies focused on word embedding for 1D, 2D, and 3D outputs with convolutional neural network (CNN). Yet, using 3D data and CNN did not perform well. Despite using CNN and 3D variants, identifying natural disaster reports remains below 80% accuracy. This research aims to improve identifying earthquakes, floods, and forest fires with 3D data and hybrid CNN long short-term memory (LSTM). The study found models with accuracies of 83.38%, 83.72%, and 89.03% for each disaster type. Hybrid CNN LSTM significantly enhanced identification compared to CNN alone, supported by statistical tests with P value less than 0.0001.
Analisis Engagement Rate Video Marketing untuk Promosi Digital Klinik Kamandaka di Instagram dan TikTok Muhammad Anis Nur Fauzi; Imam Tahyudin
Jurnal Desain Komunikasi Visual Nirmana Vol. 26 No. 2 (2026): JULY 2026
Publisher : Institute of Research and Community Outreach - Petra Christian University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.9744/nirmana.26.2.114-128

Abstract

Media sosial kini menjadi sarana utama dalam strategi promosi digital karena kemampuannya menyampaikan pesan visual secara interaktif. Klinik Kamandaka sebagai penyedia layanan kesehatan menghadapi tantangan dalam promosi digital yang sebelumnya hanya menggunakan gambar statis dengan Engagement Rate (ER) audiens yang rendah. Kondisi ini menimbulkan kebutuhan akan strategi komunikasi yang lebih menarik dan dinamis melalui video marketing. Penelitian ini bertujuan menganalisis penerapan strategi video marketing melalui platform Instagram Reels dan TikTok sebagai media promosi layanan kesehatan. Metode yang digunakan adalah kualitatif deskriptif dengan model Multimedia Development Life Cycle (MDLC) untuk menggambarkan proses pra-produksi hingga distribusi, serta analisis efektivitas menggunakan data Engagement Rate (ER). Hasil penelitian menunjukkan bahwa strategi video marketing mampu meningkatkan efektivitas promosi dengan rata-rata Engagement Rate sebesar 1,95% di Instagram dan 2,87% di TikTok. Konten dokumentasi dan edukasi menghasilkan interaksi tertinggi di Instagram, sementara konten edukasi lebih efektif di TikTok. Penerapan video berbasis multimedia terbukti memperkuat perhatian, minat, dan kepercayaan audiens terhadap layanan klinik.
ANALISIS EFEKTIVITAS VIDEO MOTION GRAPHIC SEBAGAI MEDIA PROMOSI UMKM DI KABUPATEN PEMALANG Muhamad Topan Dinar; Imam Tahyudin
Rabit : Jurnal Teknologi dan Sistem Informasi Univrab Vol 11 No 2 (2026): Juli
Publisher : LPPM Universitas Abdurrab

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36341/rabit.v11i2.8416

Abstract

Digital transformation has become a primary determinant in business communication effectiveness, yet MSMEs in the DM Kuliner area of Pemalang Regency still rely on static, conventional promotional formats. This study aims to examine the effectiveness of using motion graphics videos as a multimedia tool to improve usability perceptions and information acceptance among MSME operators. Using a One-Group Pretest-Posttest design, data were collected from 21 respondents using the System Usability Scale (SUS) instrument, which had been rigorously tested for validity and reliability. Data analysis was comprehensively conducted through descriptive statistics, the Shapiro-Wilk normality test, and hypothesis testing using the Paired Sample T-Test via SPSS. The results showed a significant increase in the overall mean SUS score from 62.26 in the pre-test phase to 69.64 in the post-test phase. Based on Acceptability Ranges mapping, this promotional medium successfully transitioned from the "Marginal Passive" category to "Acceptable" with a practical magnitude effect size (Cohen's d) of 0.66 (Medium Effect). Statistical analysis supports this finding, showing a calculated -value of -3.014 with a significance coefficient (Sig. 2-tailed) of 0.007 (), proving the increase is scientifically valid. The practical implication confirms that interactive motion graphics integration is highly effective in minimizing the digital divide and expanding market penetration for local culinary MSMEs in alignment with local government development policies.  
Klasifikasi Edibilitas Jamur Secara Otomatis Menggunakan  Algoritma Random Forest Berbasis Morfologi Muhammad Reza Pahlevi; Imam Tahyudin; Ades Tikaningsih
Jurnal Sistem Informasi Vol. 12 No. 2 (2025)
Publisher : Universitas Serang Raya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30656/jsii.v12i2.10868

Abstract

Jamur merupakan organisme yang memiliki keragaman morfologi yang tinggi, namun beberapa jenis di antaranya bersifat beracun dan membahayakan jika dikonsumsi. Kesamaan ciri fisik antara jamur yang dapat dimakan dan yang beracun sering kali menyulitkan proses identifikasi secara manual. Penelitian ini bertujuan mengembangkan sistem klasifikasi otomatis berbasis algoritma Random Forest untuk membedakan jamur edible dan poisonous, memanfaatkan seluruh 21 atribut (18 kategorikal, 3 numerik) dari dataset komprehensif Kaggle (61.069 entri). Metodologi penelitian mengikuti alur CRISP-DM yang dimodifikasi, dimulai dari pengumpulan data hingga implementasi. Tahap pra-pemrosesan data krusial dilakukan secara ekstensif, meliputi penanganan duplikasi data dan imputasi missing value (menggunakan median dan modus). Selanjutnya, transformasi label kelas (edible=0, poisonous=1) dan One-Hot Encoding diterapkan pada fitur kategorikal untuk representasi numerik yang tepat. Fitur numerik seperti cap-diameter dan stem-height dinormalisasi menggunakan Standard Scaling untuk menyeimbangkan kontribusi. Data kemudian dibagi 80:20 untuk pelatihan dan pengujian. Model Random Forest dikembangkan dengan parameter optimal (n_estimators=200, max_depth=15, class_weight="balanced") untuk efisiensi dan robustabilitas terhadap ketidakseimbangan kelas. Hasil evaluasi menunjukkan performa sangat baik dengan akurasi keseluruhan 99,34%, serta nilai precision, recall, dan f1-score yang seimbang pada 0.99 untuk kedua kelas. Analisis feature importance mengidentifikasi stem-width, stem-height, dan cap-diameter sebagai atribut paling berpengaruh. Learning curve menunjukkan stabilitas model tanpa overfitting. Implementasi pada sampel jamur baru juga mengkonfirmasi kemampuan prediksi yang konsisten, menjadikan model ini layak sebagai sistem pendukung keputusan otomatis dalam deteksi jamur beracun.
High-Accuracy Stroke Detection System Using a CBAM-ResNet18 Deep Learning Model on Brain CT Images Imam Tahyudin; R Rizal Isnanto; Anton Satria Prabuwono; Taqwa Hariguna; Eko Winarto; Nazwan Nazwan; Ades Tikaningsih; Puji Lestari; Rofik Abdul Rozak
Journal of Applied Data Sciences Vol 6, No 1: JANUARY 2025
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v6i1.569

Abstract

Stroke is a brain dysfunction that occurs suddenly as a result of local or overarching damage to the brain, lasts for at least 24 hours, and causes about 15 million deaths each year globally. Immediate medical treatment is essential to reduce the potential for further brain damage in stroke patients. Medical imaging, especially computed tomography (CT scan), plays a crucial role in the diagnosis of stroke. This study aims to develop and evaluate a deep learning architecture based on Convolutional Block Attention Module (CBAM) and ResNet18 for stroke classification in CT images. This model is designed through data preprocessing, training, and evaluation stages using a cross-validation approach. The results showed that the CBAM-ResNet18 integration resulted in a high accuracy of 95% in distinguishing stroke and non-stroke cases. The accuracy rate reached 96% for nonstroke identification (class 0) and 94% for stroke (class 1), with recall rates of 96% and 93%, respectively. Outstanding classification ability is demonstrated by an Area Under the Curve (AUC) value of 0.99. In comparison, the standard ResNet18 model shows significant fluctuations in validation loss and difficulty in generalization, with training accuracy only reaching 64-68%. On the other hand, CBAM-ResNet18 showed a significant performance improvement with a validation accuracy of 95%, a validation loss of 0.0888, and good generalization on new data. However, the limitations of the dataset and the interpretation of the results indicate the need for further validation to ensure the generalization of the model. These results show the great potential of the CBAM-ResNet18 architecture as an innovative tool in stroke diagnostic technology based on CT imaging analysis. This technology can support faster and more accurate clinical decision-making, as well as open up opportunities for further research related to the development of artificial intelligence-based systems in the medical field.
Traditional-Enhance-Mobile-Ubiquitous-Smart: Model Innovation in Higher Education Learning Style Classification Using Multidimensional and Machine Learning Methods Irfan Santiko; Tri Retnaningsih Soeprobowati; Bayu Surarso; Imam Tahyudin; Zainal Arifin Hasibuan; Ahmad Naim Che Pee
Journal of Applied Data Sciences Vol 6, No 1: JANUARY 2025
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jads.v6i1.598

Abstract

Learning achievement is undoubtedly impacted by each person's unique learning style. The assessment pattern is less focused due to the intricacy of the current components. In fact, general elements like VARK are thought to create complexity that can impair focus when combined with elements like environmental conditions, teacher effectiveness, and stakeholder policies. Although it is only ideal in specific areas, the application of supported information technology has so far yielded positive results. This essay attempts to be creative in evaluating how well students learn in higher education settings. An assessment framework that uses multidimensionality and simplifies features is the innovation that is being offered. Method, Material, and Media (3M) are the three categories into which simplification of aspects is separated. However, the Dimensions are categorized into five groups: Traditional, Enhance, Mobile, Ubiquitous, and Smart (TEMUS). Approximately 1200 respondents consisting of students and lecturers formed into a dataset in 2 types of data, namely test data and training data. The trial was conducted using 4 models, namely Random Forest, SVM, Decision Tree, and K-Nearest. The test results were interpreted in MSE, R-Square, Accuracy, Recall, Precision, and F1-Score. Based on the comparison of test results, it states that Random Forest has the most optimal results with MSE values of 0.46, R Square 0.99, Accuracy 0.86, Recall 0.86, Precision 0.87, F1 Score 0.84. Based on the results obtained, it proves that in addition to being able to carry out the classification process, the TEMUS Dimensional Framework can form a pattern of compatibility with each other, between the learning styles of Lecturers and Students. According to this TEMUS framework, teacher and student performance will be deemed suitable and effective when the 3M components are assessed from both perspectives in the same way. If not, a review will be conducted.
Graph-Enhanced Knowledge Tracing for Explainable Learner Modeling: A Systematic Literature Review Nurfaizah; R Rizal Isnanto; Imam Tahyudin
Journal of Information System and Informatics Vol 8 No 5 (2026): October
Publisher : Asosiasi Doktor Sistem Informasi Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63158/journalisi.v8i5.1844

Abstract

Knowledge Tracing (KT) has become a central approach for modeling students' evolving knowledge states, but conventional sequence-based KT models offer limited representation of relationships among skills, concepts, and assessment indicators, reducing their interpretability for monitoring mastery and learning progression. This study systematically reviews how graph-based and knowledge graph-enhanced KT models have been developed, integrated, and evaluated to support skill mastery, learning outcome monitoring, and assessment-related interpretation. A PRISMA-based Systematic Literature Review was conducted using Scopus and IEEE Xplore, covering publications indexed from 2015 to 2025. From 116 initial records, 20 studies satisfied the eligibility criteria for active graph-enhanced KT and were included in the synthesis corpus, comprising 8 core (40.0%) and 12 supporting (60.0%) studies. Six major integration families were identified, including GCN/GNN-based, Knowledge Graph/ontology-enhanced, heterogeneous or multiplex graph, attention-over-graph, Q-matrix/concept map-based KT, and graph-supported recommendation. Graph-enhanced KT can enrich learner-state representation, support skill and mastery interpretation, and strengthen explainable learner modeling, particularly among core studies. However, active graph-KT integration remains limited, and issues of construct validity, cross-domain generalization, fairness, and deployment-oriented evaluation remain underexplored.