Garuda - Garba Rujukan Digital

A Web-Based Machine Learning Approach for Standardized Precipitation Index Prediction Hadi, Ahmad Fauzi Faishal; Sinambela, Marzuki; Rachmawardani, Agustina; Trihadi, Edward
Jurnal Teknik Informatika UNIKA Santo Thomas Vol 10 No. 1 : Tahun 2025
Publisher : LPPM UNIKA Santo Thomas

Show Abstract | Download Original | Original Source | Check in Google Scholar

Accurate and user-friendly drought forecasting tools are crucial for mitigating the impact of meteorological droughts, particularly in vulnerable areas such as South Sumatra, Indonesia. This study introduces an interactive web-based application built to anticipate drought conditions by forecasting the Standardized Precipitation Index (SPI). The system relies on deep learning techniques trained using three decades of rainfall data collected from the Climatological Station in South Sumatra. In evaluating model performance, both Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) architectures were assessed. While both models delivered comparable short-term predictions, the LSTM experienced a significant decline in accuracy over extended forecasting periods (specifically at SPI-6), primarily due to overfitting. In contrast, the RNN displayed more stable and reliable results, making it the preferable model for this geographical context. Specifically, the RNN achieved a lower Mean Absolute Error (MAE) of 0.4007, a reduced Root Mean Squared Error (RMSE) of 0.4684, and a higher coefficient of determination (R²) of 0.7338. These metrics outperformed those of the LSTM, which recorded a MAE of 0.4115, an RMSE of 0.4840, and an R² of 0.7036. Such results confirm that the RNN offers a more precise and dependable fit for the station’s dataset. The web platform also effectively visualizes the model outputs, providing a dynamic and interactive 24-month forecast view that supports early warning efforts and informed decision-making for regional authorities and stakeholders.

A Web-Based Machine Learning Approach for Standardized Precipitation Index Prediction Hadi, Ahmad Fauzi Faishal; Sinambela, Marzuki; Rachmawardani, Agustina; Trihadi, Edward
Jurnal Teknik Informatika UNIKA Santo Thomas Vol 10 No. 1 : Tahun 2025
Publisher : LPPM UNIKA Santo Thomas

Show Abstract | Download Original | Original Source | Check in Google Scholar

Accurate and user-friendly drought forecasting tools are crucial for mitigating the impact of meteorological droughts, particularly in vulnerable areas such as South Sumatra, Indonesia. This study introduces an interactive web-based application built to anticipate drought conditions by forecasting the Standardized Precipitation Index (SPI). The system relies on deep learning techniques trained using three decades of rainfall data collected from the Climatological Station in South Sumatra. In evaluating model performance, both Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) architectures were assessed. While both models delivered comparable short-term predictions, the LSTM experienced a significant decline in accuracy over extended forecasting periods (specifically at SPI-6), primarily due to overfitting. In contrast, the RNN displayed more stable and reliable results, making it the preferable model for this geographical context. Specifically, the RNN achieved a lower Mean Absolute Error (MAE) of 0.4007, a reduced Root Mean Squared Error (RMSE) of 0.4684, and a higher coefficient of determination (R²) of 0.7338. These metrics outperformed those of the LSTM, which recorded a MAE of 0.4115, an RMSE of 0.4840, and an R² of 0.7036. Such results confirm that the RNN offers a more precise and dependable fit for the station’s dataset. The web platform also effectively visualizes the model outputs, providing a dynamic and interactive 24-month forecast view that supports early warning efforts and informed decision-making for regional authorities and stakeholders.

TRANSFORMASI PUBLIKASI STMKG DIGITAL: PENINGKATKAN SUMBER DAYA MANUSIA UNGGUL DAN PERCEPATAN AKREDITASI INSTITUSI Sinambela, Marzuki; Hidayat, Nur; Adi, Suko Prayitno; Sulistya, Widada; Sudarisman, Maman; Riama, Nelly Florida
Majalah Ilmiah METHODA Vol. 13 No. 2 (2023): Majalah Ilmiah METHODA
Publisher : Universitas Methodist Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.46880/methoda.Vol13No2.pp207-216

Improving excellent human resources is an initiative that aims to develop and strengthen individual and collective potential within the BMKG (Meteorology, Climatology and Geophysics Agency) organization to make it a leading global player in the fields of meteorology, climatology, geophysics and instrumentation. Excellent human resources in higher education institutions have the potential to conduct quality research and produce quality scientific publications. In this action of change, the integration of the STMKG Digital publication service program, both updating the E-Journal and building the STMKG PRESS publishing media as a single digital-based and indexed account, has been successfully carried out and is the key to the realization of a comprehensive publication information system. The integration of publication services is aimed at improving efficiency and effectiveness in both indexing and digital documents. This transformation will encourage teams involved in this change action plan to collaborate more, effective communication, and writing literacy. The results of this change action are expected to be useful for STMKG's internal interests, namely to facilitate the accreditation preparation process, academic data collection, and the accreditation assessment simulation process. The benefits for BMKG are the implementation of the BMKG 2022-2024 strategic plan and the improvement of superior human resources towards 500 Doctorates and BMKG Global Player.

VISUAL ANALYSIS OF LOCAL EARTHQUAKE IN NORTH TAPANULI BASED ON DATA SCIENCE Sinambela, Marzuki; Darnila, Eva
METHOMIKA: Jurnal Manajemen Informatika & Komputerisasi Akuntansi Vol. 7 No. 2 (2023): METHOMIKA: Jurnal Manajemen Informatika & Komputersisasi Akuntansi
Publisher : Universitas Methodist Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.46880/jmika.Vol7No2.pp363-367

Earthquakes are natural phenomena that occur when the Earth's tectonic plates move and release energy. Big Data's emergent epistemological and research paradigms, as well as data science, an increasingly integrated field of data research, are opening up new opportunities. Visualizing earthquake data is all about understanding earthquake characteristics such as size, location and depth. The result show that September was the quietest month in terms of earthquakes, and in this graph we can see the number of earthquakes for each month in 2022. The month of October is the one that has the highest number of earthquakes. We can see the average depth and magnitude of each year on the bubble chart. In addition, the size and color of the bubbles indicate the number of earthquakes that month. In general, most of the earthquakes occurred in the shallow earthquake range and the 1.8-3.85 magnitude range.

Mitigasi Bencana Gempa Bumi dengan Integrasi Analisis Geofisika dan Data Mining Yudha, I Putu Putra Wira Sarwa; Sinambela, Marzuki
Geosfera: Jurnal Penelitian Geografi Vol 3, No 2 (2024): Geosfera: Jurnal Penelitian Geografi
Publisher : Universitas Negeri Gorontalo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37905/geojpg.v3i2.24971

Kabupaten Cianjur merupakan salah satu kabupaten di Provinsi Jawa Barat yang rentan terhadap bencana gempa bumi karena dilewati sesar Cimandiri. Pada akhir tahun 2022, misalnya, telah terjadi insiden gempa bumi besar yang menghantam Kabupaten Cianjur. Penelitian ini bertujuan untuk menyelidiki aktivitas gempa bumi di Kabupaten Cianjur, Jawa Barat, Indonesia, dengan mengintegrasikan analisis multimetode Geofisika dan data resmi dari Portal Satu Data Indonesia. Aspek yang akan diteliti meliputi pola distribusi, frekuensi kejadian, karakteristik gempa bumi, dan faktor-faktor yang mempengaruhi aktivitas seismik di wilayah tersebut. Metode penelitian mencakup pengambilan data episenter dan hiposenter gempa bumi hasil relokasi oleh BMKG, data hasil pengamatan gempa bumi dari BMKG, serta data jumlah kejadian bencana alam yang diunduh dari Portal Satu Data Indonesia. Data hasil pengamatan dari BMKG akan dipadukan dengan data mining pada dataset sumber. Analisis data bertujuan untuk mengidentifikasi sumber, penyebab, dan karakteristik gempa, serta mengelola informasi dari data yang besar menjadi ringkas dan mudah dipahami. Sebagai referensi, penelitian ini akan menyertakan kajian pustaka dari penelitian-penelitian yang membahas kasus gempa bumi dari seluruh dunia. Diharapkan, informasi yang komprehensif tentang karakteristik gempa bumi di wilayah Cianjur ini dapat berkontribusi dalam membangun mitigasi bencana yang efektif.

Timeseries forecasting for Local Average Temperature in Northern Sumatera Using Long Short-Term Memory Model Sinambela, Marzuki; Sudarisman, Maman; Munawar, Munawar
JUKI : Jurnal Komputer dan Informatika Vol. 5 No. 2 (2023): JUKI : Jurnal Komputer dan Informatika, Edisi Nopember 2023
Publisher : Yayasan Kita Menulis

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.53842/juki.v5i2.385

For better management and planning of water resources in a basin, it is important to understand trends and predict average temperature as one of the parameters of weather and climate data. The study of weather trends using normal and local annual average temperature, comparison and observation. In this study, we will analyse the local and normal average temperature data in the city of Medan, based on the observation station in situ. The main objective of this study is to compare the normal temperature with the local station and to predict the temperature data in the city of Medan, North Sumatra by using the long term short term memory model. Based on the result of normal data science of exploring temperature with local temperature correlation, we got the display of training curve, residual plot and the scatter plot are shown using these codes. The good performance of Kualanamu and better than Deliserdang station had MSE value 0.01 and R2 value 0.98, close to zero represents better prediction quality.

Spatiotemporal Dynamics of Seismic Activity in the Toba Caldera based on DBSCAN Clustering Algorithms Sinambela, Marzuki; Purwantiningsih, Purwantiningsih; Anita, Febria; Hartoyo, Puji; Mutanto, Ari
JUKI : Jurnal Komputer dan Informatika Vol. 7 No. 2 (2025): JUKI : Jurnal Komputer dan Informatika, Edisi Nopember 2025
Publisher : Yayasan Kita Menulis

Show Abstract | Download Original | Original Source | Check in Google Scholar

Salah satu sistem vulkanik terpenting di dunia, Kaldera Toba, memiliki lingkungan seismotektonik yang kompleks akibat interaksi antara tektonik regional, proses vulkanik, dan dinamika kaldera. Kinerja algoritma Density-Based Spatial Clustering of Applications with Noise (DBSCAN) menjadi fokus utama penelitian ini, yang menggunakan pendekatan pembelajaran mesin tanpa pengawasan untuk mengevaluasi aktivitas seismik di Toba Geopark antara tahun 2019 dan 2022. Sebagai fitur masukan, parameter gempa seperti magnitudo, percepatan tanah maksimum, kedalaman hiposentrum, dan posisi geografis digunakan. Indeks Davies-Bouldin, indeks Calinski-Harabasz, dan koefisien siluet digunakan untuk menilai kinerja pengelompokan. Berbeda dengan kinerja yang diamati di lokasi yang dipengaruhi tektonik, hasil menunjukkan bahwa DBSCAN sangat efektif di lingkungan vulkanik, mencapai skor siluet 0,679 dan indeks Davies-Bouldin 0,404. Sifat diskrit dan terkendali struktur dari seismisitas vulkanik tercermin dalam identifikasi DBSCAN terhadap enam kluster seismik kompak yang terkait dengan struktur kaldera unik dan klasifikasinya terhadap 91,48% peristiwa sebagai noise. Korelasi yang kuat antara kluster yang terdeteksi dan karakteristik vulkanik yang dikenal, seperti tepi kaldera, kompleks vulkanik pusat, dan sistem patahan yang berinteraksi, terungkap melalui analisis spasial. Hasil ini menunjukkan keefektifan algoritma klustering yang spesifik lingkungan dan membantu dalam pengembangan metode berbasis pembelajaran mesin untuk penilaian risiko dan pemantauan seismik.

Peningkatan Literasi Kebencanaan untuk Optimalisasi Informasi Peringatan Dini Bencana Geo-Hidrometeorologi Munawar, Munawar; Haryanto, Yosafat Donni; Abigael, Febby Debora; Muftareza, Arfany Dimas; Muthahhari, Ilham; Mardiyansyah, Adji; Sinambela, Marzuki
JPM: Jurnal Pengabdian Masyarakat Vol. 6 No. 3 (2026): January 2026
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/jpm.v6i3.2629

Tangerang City is one of the regions in Banten Province that is vulnerable to geo-hydrometeorological disasters and has a relatively large population with rapid urban growth, which can increase the risk of such disasters. Therefore, information, understanding, and actions regarding geo-hydrometeorological disasters are required. Environmental damage and land use change, such as deforestation, urbanization, river sedimentation, and land function conversion, are significant issues. This activity aims to enhance disaster literacy based on Meteorology, Climatology, Geophysics, and Instrumentation (MKGI) in Tangerang City, scheduled for Monday, July 21, 2025, at SMA Al-Husna, Tangerang City, with a total of 67 participants. Through interactive methods that include material presentation, question and answer sessions, quizzes, and questionnaire filling, participants are provided with a comprehensive understanding of geo-hydrometeorological disasters, mitigation strategies, and the utilization of technology in early warning systems. One of the innovations of this activity is the development of a weather forecast information product based on Telegram bot for the area of Tanah Tinggi Village, which provides weather forecast information automatically using open data from BMKG. Evaluation results show that participants had a good understanding of the socialization material, especially in the topics of information dissemination and geohydrometeorological disasters. The questionnaire index score reached 87.60%, falling into the category of "Strongly Agree". The developed Telegram product successfully presents real-time weather information every 3 hours in a structured and easily accessible manner. This activity proves that a literacy approach based on MKGI and technology can have a positive impact in raising awareness and improving preparedness against geo-hydrometeorological disaster risks.

Title

Found 18 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 18 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 18 Documents
Search