Claim Missing Document
Check
Articles

Found 23 Documents
Search

Analisis Forensik Digital Terhadap Perdagangan Data Pribadi Di Dark Web Menggunakan Osint & Threat Intelligence Dalimunthe, Ahmad Al Qodri Azizi; Azhari, Mulkan
Jurnal Komputer Teknologi Informasi Sistem Komputer (JUKTISI) Vol. 4 No. 1 (2025): Juni 2025
Publisher : LKP KARYA PRIMA KURSUS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62712/juktisi.v4i1.400

Abstract

Kebocoran data pribadi yang diperjualbelikan di Dark Web menjadi isu yang semakin mengkhawatirkan, terutama setelah kasus yang menimpa Kementerian Pendidikan, Kebudayaan, Riset, dan Teknologi (Kemendikbudristek) pada tahun 2024. Penelitian ini bertujuan untuk menganalisis pola perdagangan data pribadi di Dark Web dengan pendekatan forensik digital yang didukung oleh metode Open Source Intelligence (OSINT) dan Threat Intelligence. Penelitian dilakukan dengan studi kasus terhadap data yang dibagikan oleh akun “grepcn” di forum LeakBase dan disebarkan ulang oleh akun “knox” di DarkForums. Proses investigasi dilakukan melalui pemantauan pasif, analisis struktur data dengan tools seperti Python dan NetworkX, serta validasi email menggunakan platform OSINT seperti HaveIBeenPwned dan IntelX. Hasilnya menunjukkan bahwa data pribadi diperjualbelikan dalam format SQL dan disembunyikan di balik sistem berbayar menggunakan mata uang kripto. Sebagian besar data yang dianalisis terbukti valid dan pernah mengalami kebocoran. Penelitian ini menunjukkan bahwa pendekatan gabungan OSINT dan Threat Intelligence dapat digunakan secara efektif untuk mendeteksi dan menganalisis aktivitas perdagangan data pribadi di Dark Web, serta memberikan gambaran awal mengenai ancaman siber yang semakin berkembang.
Analisis Kinerja Algoritma K-Means dan K-Medoids dalam Pengelompokan Penerima Bantuan Sosial di Kelurahan Terjun Siti Khairun Nisa; Mulkan Azhari
Jurnal ilmiah Sistem Informasi dan Ilmu Komputer Vol. 5 No. 2 (2025): Juli : Jurnal ilmiah Sistem Informasi dan Ilmu Komputer
Publisher : Lembaga Pengembangan Kinerja Dosen

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55606/juisik.v5i2.1293

Abstract

The advancement of digital technology has improved data management, including in the distribution of social assistance. However, the large volume of data and the similarity of community characteristics often hinder the manual determination of aid recipients. This study analyzes the performance of two clustering algorithms, K-Means and K-Medoids, in grouping social assistance recipients in Kelurahan Terjun. Using a quantitative approach and data mining techniques based on clustering. The data is divided into three groups: Eligible, Not Eligible, and Requires Validation. The results show that although both algorithms produce similar clustering patterns, K-Medoids demonstrates better performance in cluster distribution and visualization. Cluster visualization using PCA indicates that K-Medoids forms clearer cluster boundaries and more balanced data distribution compared to K-Means. It can be concluded that K-Medoids outperforms in clustering social assistance recipient data and can serve as a more efficient alternative for targeted aid distribution.
Application of Data Mining to Determine the Performance of Family Planning Field Officers (PLKB) Using the C4.5 Algorithm Nasution, Perdinal; Azhari, Mulkan
Hanif Journal of Information Systems Vol. 3 No. 1 (2025): August Edition
Publisher : Ilmu Bersama Center

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.56211/hanif.v3i1.52

Abstract

The effectiveness of family planning programs is closely related to the performance of Family Planning Field Officers (PLKB). Conventional performance evaluation methods often rely on manual assessments, which may lead to subjectivity and inconsistency. To overcome this issue, data mining techniques can be applied to analyze performance data systematically and objectively. This study employs the C4.5 decision tree algorithm to classify and evaluate the performance of PLKB. The dataset used in this research includes several indicators, such as service coverage, counseling frequency, reporting accuracy, and community participation. Prior to model construction, data preprocessing was performed to handle missing values and normalize attributes. The model performance was evaluated using accuracy, precision, recall, and F-measure. The findings indicate that the C4.5 algorithm successfully classified PLKB performance into three categories: high, medium, and low. The model achieved an accuracy of [insert % if available], demonstrating its effectiveness in identifying key determinants of officer performance. Moreover, the decision tree generated interpretable rules that highlight the most influential attributes affecting PLKB performance. The application of data mining using the C4.5 algorithm provides an objective and efficient method for evaluating PLKB performance. This approach not only enhances decision-making for supervision and training but also contributes to the improvement of family planning program implementation. Future research is suggested to compare the C4.5 algorithm with other classification methods to achieve higher accuracy and generalizability.
Analisis Perbandingan Metode LSTM dan BiLSTM untuk Prediksi Harga Saham Menggunakan Alpha Vantage Rifqi Yafik; Mulkan Azhari
Jurnal Komputer Teknologi Informasi Sistem Komputer (JUKTISI) Vol. 4 No. 3 (2026): Februari 2026
Publisher : LKP KARYA PRIMA KURSUS

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62712/juktisi.v4i3.650

Abstract

Pasar saham Indonesia mengalami pertumbuhan signifikan, namun fluktuasi harga yang tinggi tetap menjadi tantangan utama bagi investor. Penelitian ini bertujuan membandingkan performa dua model deep learning, Long Short-Term Memory (LSTM) dan Bidirectional LSTM (BiLSTM), untuk memprediksi harga penutupan harian dua saham blue chip, PT Bank Central Asia Tbk (BBCA) dan PT Telkom Indonesia Tbk (TLKM). Data historis dari Januari 2019 hingga Desember 2023 diperoleh melalui Alpha Vantage API. Proses penelitian mencakup normalisasi data dengan MinMaxScaler dan pembentukan sliding window untuk pemodelan deret waktu. Model LSTM dan BiLSTM dilatih menggunakan TensorFlow-Keras, dengan optimasi hyperparameter melalui metode Grid Search yang menguji kombinasi units, batch size, epochs, dan dropout rate. Hasil eksperimen menunjukkan bahwa model BiLSTM memberikan akurasi prediksi yang lebih tinggi dibandingkan LSTM pada kedua saham. Untuk saham BBCA, BiLSTM mencatat RMSE sebesar 0.0178, lebih baik dari LSTM yang mencatat RMSE 0.0180. Begitu pula pada saham TLKM, BiLSTM mencapai RMSE 0.0172, mengungguli LSTM dengan RMSE 0.0199. Keunggulan BiLSTM disebabkan kemampuannya memproses data secara dua arah, yang memungkinkan model menangkap pola kontekstual dan titik pembalikan tren dengan lebih baik. Penelitian ini berkontribusi pada pengembangan model prediksi saham yang lebih akurat dan sistematis bagi peneliti dan praktisi di pasar modal.
Implementation of The Sales and Purchase Program Application Using The Rapid Application Development Model Web – Based Ichsan, Aulia; Al-Khowarizmi, Al-Khowarizmi; Azhari, Mulkan
Tsabit Journal of Computer Science Vol. 1 No. 1 (2024): June Edition
Publisher : Ilmu Bersama Center

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.56211/tsabit21

Abstract

The development of information technology is currently developing rapidly and rapidly, which is supported by one of the means, namely computers. Of course, computers that are equipped with certain applications are used to help make human work easier in managing data for an organization or company so that they get accurate results that meet their needs. The results of observations that have been made show that sales and purchasing activities still use manual systems, one of which is in clothing stores. Starting from processing goods data, difficulties checking stock, purchasing transactions, sales transactions, as well as storing other data related to all types of activities, which can cause losses for shop owners, errors in recording and inaccurate reports being made. Judging from the large number of transactions carried out at clothing stores, a faster and more accurate information system is needed. Therefore, the author created a computerized program design using the Microsoft Visual Basic.net programming language and MySQL database, so that information and activities that occur can be carried out quickly and accurately. The method used in designing this program uses the Rapid Application Development (RAD) model. This RAD model is an adaptation of the high-speed version of the waterfall model for the development of each software component. The results achieved from discussing this theme are in the form of ready-to-use sales and purchasing program applications. In this case, the use of program applications is the best solution to solve existing problems, and by using program applications an effective and efficient activity can be achieved in supporting activities, especially for handling sales and purchasing problems.
The Aplikasi Model Text Area Based Image Selective Encryption Menggunakan YoloV3, Arnold's Catmap dan AES Pada Pengamanan Konten Teks Pada Citra Digital Riza, Ferdy; Azhari, Mulkan; Zulherry, Andi
JURNAL RISET KOMPUTER (JURIKOM) Vol. 12 No. 5 (2025): Oktober 2025
Publisher : Universitas Budi Darma

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30865/jurikom.v12i5.9261

Abstract

sensitive content in digital images. This research proposes a selective encryption model based on text area detection in digital images, integrating object detection using You Only Look Once version 3 (YOLOv3), Arnold's Cat Map transformation, and the Advanced Encryption Standard (AES) algorithm. The model automatically identifies and selects areas containing text in the image using YOLOv3, applies Arnold's Cat Map for spatial disorganization, and then encrypts the transformed result with AES to ensure data security. System performance is evaluated through visual quality analysis using PSNR (Peak Signal-to-Noise Ratio) and SSIM (Structural Similarity Index) parameters, as well as encryption and decryption processing time. The test results show that this approach can maintain the integrity of non-text areas while providing strong protection for sensitive text areas without compromising efficiency or overall visual quality. This model has the potential to be applied in the context of securing digital documents, visual identities, and other sensitive data in images.
Implementasi Regresi Linear Berganda dalam Forecast Penjualan pada CV. Surya Kencana Sembako Berbasis website Ilham; Mulkan Azhari
Jurnal Teknologi Informasi (JUTECH) Vol 6 No 2 (2025): JUTECH: Jurnal Teknologi Informasi
Publisher : ITB Ahmad Dahlan Jakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32546/jutech.v6i2.3166

Abstract

Perkiraan penjualan merupakan aspek penting dalam pengambilan keputusan bisnis, terutama dalam industri sembako yang memiliki permintaan tinggi dan fluktuatif. CV. Surya Kencana Sembako sebagai perusahaan distribusi barang kebutuhan pokok menghadapi tantangan dalam memprediksi jumlah penjualan secara akurat. Penelitian ini bertujuan untuk mengimplementasikan metode Regresi Linear Berganda dalam meramalkan penjualan berdasarkan beberapa variabel yang memengaruhi, seperti harga produk, jumlah promosi, dan musim penjualan. Sistem dirancang dalam bentuk aplikasi berbasis website agar memudahkan pengguna dalam menginput data dan memperoleh hasil prediksi secara real-time. Hasil dari implementasi menunjukkan bahwa model regresi mampu memberikan prediksi yang cukup akurat dengan nilai koefisien determinasi (R²) yang tinggi. Dengan adanya sistem ini, diharapkan CV. Surya Kencana Sembako dapat meningkatkan efisiensi dalam pengelolaan stok dan strategi penjualannya.
Clustering Jenis Sayuran Di Daerah Desa Sempa Jaya Dengan Algoritma Gausian Mixture Model Syafik, Hafizan; Azhari, Mulkan
Portal Riset dan Inovasi Sistem Perangkat Lunak Vol. 3 No. 4 (2025): Artikel Penelitian
Publisher : SoraTekno Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.59696/prinsip.v3i4.188

Abstract

Sektor pertanian di Desa Sempajaya, Kabupaten Karo, memiliki potensi besar dalam menghasilkan berbagai jenis sayuran yang menjadi sumber utama pemenuhan kebutuhan gizi masyarakat sekaligus penyokong perekonomian lokal. Namun, pengelolaan lahan dan pemetaan jenis sayuran unggulan masih menghadapi kendala karena belum adanya sistem pengelompokan data yang akurat. Penelitian ini bertujuan untuk mengelompokkan jenis sayuran di Desa Sempajaya menggunakan algoritma Gaussian Mixture Model (GMM) sebagai metode clustering yang mampu menangani distribusi data yang kompleks. Metode penelitian dilakukan melalui beberapa tahap, yaitu pengumpulan data (observasi, wawancara, dan dokumentasi), pra-pemrosesan data, analisis faktor menggunakan diagram Fishbone untuk menentukan atribut relevan, serta implementasi algoritma GMM dengan pendekatan Expectation-Maximization (EM). Data yang digunakan mencakup enam komoditas utama, yaitu cabai, tomat, sawi, wortel, terung, dan buncis, dengan variabel meliputi ukuran, berat, warna, serta luas lahan. Hasil penelitian menunjukkan bahwa algoritma GMM berhasil mengelompokkan data sayuran ke dalam tiga kategori kluster produksi, yaitu rendah, sedang, dan tinggi, dengan visualisasi hasil clustering yang lebih representatif dibandingkan metode konvensional. Sistem ini mampu memberikan informasi potensi sayuran unggulan di setiap wilayah Desa Sempajaya, yang dapat dimanfaatkan oleh petani, masyarakat, maupun pemerintah daerah dalam pengambilan keputusan strategis terkait diversifikasi, distribusi, dan pengembangan pertanian berkelanjutan.
Development of Virtual Reality-Based Computer Assembly Simulation Learning Media Prastia, Ilham; Azhari, Mulkan
Hanif Journal of Information Systems Vol. 3 No. 1 (2025): August Edition
Publisher : Ilmu Bersama Center

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.56211/hanif.v3i1.66

Abstract

The development of Virtual Reality (VR) technology provides great opportunities in creating interactive and immersive learning media that can simulate hands-on practice more realistically. In computer assembly learning at vocational schools, limited availability of laboratory equipment often becomes a major obstacle, resulting in students not gaining optimal direct practice experience. This study aims to develop a Virtual Reality-based computer assembly learning simulation as an interactive, safe, and engaging alternative learning tool. The research employed a Research and Development (R&D) method using the ADDIE model, consisting of the stages of analysis, design, development, implementation, and evaluation. Computer component assets were modeled using Blender 3D and then integrated into Unity to build an interactive VR-based simulation. The testing phase involved Black Box Testing and Application Testing with 10 respondents, consisting of 7 vocational students and 3 alumni from the Computer and Network Engineering major. The results show that all interactive features performed according to the expected scenarios, and the feasibility assessment through Application Testing achieved a score of 87.2%, indicating that the simulation is suitable, easy to use, and effective in improving students’ understanding of computer assembly procedures. Additionally, the VR media was considered to provide a more realistic learning experience, reduce the potential for errors during real practice, and increase student engagement throughout the learning process. Therefore, this VR-based learning media can serve as a solution to laboratory limitations and a foundation for further development of VR-based practical learning materials in the field of Computer and Network Engineering.
Analisis dan Implementasi Algoritma Apriori dalam Memprediksi Banjir di Kota Medan pada Bendungan Lau Simeme Zahra Apriyani Hakim Nasution; Mulkan Azhari
Journal of Science, Technology, and Innovation Vol 1 No 3 (2026): : April: Inventa: Journal of Science, Technology, and Innovation
Publisher : CV SCRIPTA INTELEKTUAL MANDIRI

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.65310/4cn1ym72

Abstract

This study aims to analyze and implement the Apriori algorithm to predict flood potential at the Lau Simeme Dam in Medan City by identifying association patterns among rainfall, water discharge, and water level parameters based on daily hydrological data. The main challenge of this study lies in the limitations of conventional mitigation systems, which are not yet capable of systematically and adaptively interpreting multivariate environmental relationships. The research method employs an empirical data mining-based approach, involving data preprocessing, numerical transformation into transactional data, frequent item set formation, and association rule derivation using minimum support and confidence parameters. The system was developed using Python and MySQL to support the operational analysis and visualization of prediction results. The results show that the Apriori algorithm is capable of generating consistent association patterns between heavy rainfall, increased water discharge, and flood alert status with an accuracy of 97.81%, precision of 100%, and recall of 97.81%. These findings indicate that association rule-based models possess interpretive and predictive capabilities relevant to supporting flood mitigation based on hydrological data.