Article Per Year (5 Year)
p-Index From 2021 - 2026
0.562
P-Index
This Author published in this journals
All Journal AdMathEduSt: Jurnal Ilmiah Mahasiswa Pendidikan Matematika Jurnal Sistem Komputer dan Informatika (JSON) Newton-Maxwell Journal of Physics
Yosik Norman
Unknown Affiliation
Astronomy Computer Science & IT Control & Systems Engineering Earth & Planetary Sciences Electrical & Electronics Engineering Mathematics Mechanical Engineering Physics

Published : 3 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 3 Documents
Search

 1 
PERFORMANCE COMPARISON OF RANDOM FOREST, DECISION TREE, AND EXTRA TREES MODELS FOR RAINFALL PREDICTION IN JAKARTA Syahrin, Khairummin Alfi; Tiara Emanuella Disera; Juang Merdeka; Alamin, Mirza Virgiansah; Yosik Norman
AdMathEduSt: Jurnal Ilmiah Mahasiswa Pendidikan Matematika Vol. 13 No. 1 (2026)
Publisher : Universitas Ahmad Dahlan

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

Abstract

Accurate rainfall prediction is important for weather monitoring and flood management in urban areas. This study evaluates the performance of three decision tree-based models, Random Forest, Extra Trees, and Decision Tree, for predicting daily rainfall in Jakarta using data from the BMKG Tanjung Priok observation station for the year 2025. The dataset, expressed in millimeters per day, was preprocessed to handle missing values and ensure consistency, and analyzed using the PyCaret library in the Jupyter Notebook environment. Model training was conducted with optimized hyperparameters, and performance was assessed using MASE, RMSSE, MAE, RMSE, SMAPE, and R². All models produced similar overall trends, although the Extra Trees model showed slightly higher fluctuations. Comparative evaluation indicated that the Random Forest model achieved the best performance, with MASE of 0.7925, RMSSE of 0.6373, MAE of 9.51 millimeters, RMSE of 14.06 millimeters, SMAPE of 1.8287, and R² of -0.8227, demonstrating superior accuracy in capturing rainfall patterns. These results suggest that Random Forest is the most suitable model for daily rainfall forecasting in Jakarta, providing reliable predictions that can support meteorologists and policymakers in improving forecast accuracy and planning.
Perbandingan Performa Multi-Algoritma Machine Learning dengan Dua Strategi Validasi pada Klasifikasi Curah Hujan I Dewa Gede Loka Maheswara; Arya Zaki Ramadhan; Rica Azzura Maldina; Muhammad Fany Nurwibowo; Yosik Norman
Jurnal Sistem Komputer dan Informatika (JSON) Vol. 7 No. 3 (2026): Maret 2026
Publisher : Universitas Budi Darma

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30865/json.v7i3.9475

Abstract

Prediksi curah hujan yang akurat masih menjadi tantangan karena kompleksitas proses atmosfer serta dampaknya terhadap berbagai sektor. Performa algoritma machine learning dalam klasifikasi curah hujan sangat dipengaruhi oleh karakteristik data dan metode validasi, sehingga diperlukan evaluasi komparatif untuk menentukan model yang paling sesuai pada konteks lokal. Penelitian ini bertujuan membandingkan performa lima model machine learning, yaitu Random Forest, XGBoost, Support Vector Machine, K-Nearest Neighbor, dan Decision Tree dalam klasifikasi curah hujan di Kabupaten Tapanuli Tengah menggunakan data observasi harian periode 2015–2024 sebanyak 32.796 data yang diperoleh dari Stasiun Meteorologi FL Tobing. Evaluasi dilakukan melalui skema pembagian data dan 10-cross fold validation dengan metrik precision, recall, dan f1-score. Hasil penelitian menunjukkan bahwa Random Forest secara konsisten memberikan performa terbaik pada kedua skema validasi dengan f1-score sebesar 62% dan 63%, lebih stabil dibandingkan model lainnya pada kondisi distribusi kelas yang tidak seimbang. Temuan ini menunjukkan bahwa pendekatan ensemble lebih adaptif dalam menangkap hubungan nonlinier parameter meteorologi serta memberikan dasar metodologis dalam pemilihan model klasifikasi curah hujan untuk mendukung mitigasi bencana hidrometeorologi.
Prediksi Trajektori dan Intensitas Siklon Tropis Menggunakan Pendekatan Multi-Task Learning Berbasis Recurrent Neural Network Syahid, Wisnu; Putu Aldi Tusan Pratama; Muhammad Nur Rizqi; Yosik Norman
Newton-Maxwell Journal of Physics Vol. 7 No. 1: April 2026
Publisher : UNIB Press

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33369/nmj.v7i1.47955

Abstract

The limited ability of Numerical Weather Prediction (NWP) models to capture nonlinear dynamics and atmospheric uncertainty remains a major challenge in improving tropical cyclone forecasts, particularly over the eastern Indian Ocean. This study evaluates a Multi-Task Learning approach based on several Recurrent Neural Network (RNN) variants, namely LSTM, BiLSTM, GRU, and BiGRU, to simultaneously predict three key cyclone components: position (latitude and longitude), wind intensity, and cyclone category. Historical IBTrACS data from 2000 to 2025 with a 3-hour temporal resolution are used as model input, employing 48-hour sequences to forecast cyclone conditions at lead times of 12, 24, 48, and 72 hours. The results show that all models achieve stable convergence during training. At a 12-hour lead time, the BiLSTM model delivers the best performance, with a mean position error of 83.53 km and a Hit Rate of 0.966, outperforming the other models. For longer lead times (24–72 hours), the BiGRU model demonstrates the most stable positional accuracy, exhibiting the lowest error degradation as the forecast horizon increases. In addition, wind intensity predictions remain robust, with a Mean Absolute Error (MAE) below 4.6 knots up to 72 hours. These findings highlight the potential of multi-output RNN-based models to support more adaptive and efficient tropical cyclone forecasting systems.
Co-Authors Alamin, Mirza Virgiansah Arya Zaki Ramadhan I Dewa Gede Loka Maheswara Juang Merdeka Muhammad Fany Nurwibowo Muhammad Nur Rizqi Putu Aldi Tusan Pratama Rica Azzura Maldina Syahid, Wisnu Syahrin, Khairummin Alfi Tiara Emanuella Disera