Claim Missing Document
Check
Articles

Found 4 Documents
Search

Enhancing El Niño-Southern oscillation prediction using an attention-based sequence-to-sequence architecture Setiawan, Karli Eka; Fredyan, Renaldy; Alam, Islam Nur
International Journal of Electrical and Computer Engineering (IJECE) Vol 14, No 6: December 2024
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijece.v14i6.pp7057-7066

Abstract

The ability to accurately predict the EI Nino-Southern oscillation (ENSO) is essential for seasonal climate forecasting. Monitoring the Pacific Ocean's surface temperature has many benefits for human life, including a better understanding of climate and weather, the ability to predict summer and winter, the ability to manage natural resources, serving as a reference for maritime transportation and navigation needs, serving as a reference for climate change monitoring needs, and even serving as a renewable energy source by utilizing high sea surface temperatures. This study introduces a deep learning (DL) model with AttentionSeq2Luong model as our proposed model to the ENSO research community. The present study showcases the capability of our proposed model to effectively forecast the forthcoming monthly average Nino index compared to the baseline seq2seq architecture model. For the dataset, this study utilized monthly observations of Nino 12, Nino 3, Nino 34, and Nino 4 between January 1870 and August 2022. The brief result of our experiment was that applying Luong Attention in the seq2seq model reduced the RMSE error by around 0.03494, 0.04635, 0.03853, and 0.03892 for forecasting Nino 12, Nino 3, Nino 34, and Nino 4, respectively.
Leveraging Support Vector Machines and Ensemble Learning for Early Diabetes Risk Assessment: A Comparative Study Shiddiqi, Hafizh Ash; Setiawan, Karli Eka; Fredyan, Renaldy
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 1 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i1.12846

Abstract

Currently, diabetes is a hidden, serious threat to human lifestyles through daily food and drink, which has become a formidable global health challenge. As a contribution, this study suggests a way to use machine learning to find people with diabetes by looking at certain health parameters. It does this by using different Support Vector Machine (SVM)-based models, such as different SVMs with different kernels, such as linear, polynomial, radial basis function, and sigmoid kernels; different ensemble bagging with SVM; and different ensemble stacking with various SVM models. The findings demonstrated that utilizing a single SVM model with a linear kernel, ensemble bagging with a linear SVM, and ensemble stacking with different SVM models yielded the most accurate results, achieving 95% accuracy in both diabetes presence and absence. This lends credence to the idea that the incorporation of a linear kernel has the potential to improve the accuracy of determining whether or not diabetic illness is present.
Enhancing spatiotemporal weather forecasting accuracy with 3D convolutional kernel through the sequence to sequence model Fredyan, Renaldy; Setiawan, Karli Eka; Minor, Kelvin Asclepius
International Journal of Electrical and Computer Engineering (IJECE) Vol 15, No 2: April 2025
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijece.v15i2.pp2022-2030

Abstract

Accurate weather forecasting is important when dealing with various sectors, such as retail, agriculture, and aviation, especially during extreme weather events like heat waves, droughts, and storms to prevent disaster impact. Traditional methods rely on complex, physics-based models to predict the Earth's stochastic systems. However, some technological advancements and the availability of extensive satellite data from beyond Earth have enhanced meteorological predictions and sent them to Earth's antennae. Deep learning models using this historical data show promise in improving forecast accuracy to enhance how models learn the data pattern. This study introduces a novel architecture, convolutional sequence to sequence (ConvSeq2Seq) network, which employs 3D convolutional neural networks (CNN) to address the challenges of spatiotemporal forecasting. Unlike recurrent neural network (RNN)--based models, which are time-consuming due to sequential processing, 3D CNNs capture spatial context more efficiently. ConvSeq2Seq overcomes the limitations of traditional CNN models by ensuring causal constraints and generating flexible length output sequences. Our experimental results demonstrate that ConvSeq2Seq outperforms traditional and modern RNN-based architectures in both prediction accuracy and time efficiency, leveraging historical meteorological data to provide a robust solution for weather forecasting applications. The proposed architecture outperforms the previous method, giving new insight when dealing with spatiotemporal with high density.
Advancing Indonesian Audio Emotion Classification: A Comparative Study Using IndoWaveSentiment Majiid, Muhammad Rizki Nur; Setiawan, Karli Eka; Pamungkas, Prayoga Yudha; Annas, Taufiq; Setiawan, Nicholas Lorenzo
Engineering, MAthematics and Computer Science Journal (EMACS) Vol. 7 No. 2 (2025): EMACS
Publisher : Bina Nusantara University

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21512/emacsjournal.v7i2.13415

Abstract

This study addresses the critical gap in Indonesian Speech Emotion Recognition (SER) by evaluating machine learning models on the IndoWaveSentiment dataset, a novel corpus of 300 high-fidelity recordings capturing five emotions (neutral, happy, surprised, disgusted, disappointed) from native speakers. The research aims to identify optimal classification techniques and acoustic features for Indonesian SER, given the language’s unique linguistic characteristics and the scarcity of annotated resources. Six models, Logistic Regression, KNN, Gradient Boosting, Random Forest, Naive Bayes, and SVC, were trained on 45 acoustic features, including spectral contrast, MFCCs, and zero crossing rate, extracted using Librosa. Results demonstrated Random Forest as the top performer (90% accuracy), followed by Gradient Boosting (85%) and Logistic Regression (75%), with spectral contrast (contrast2, contrast7) and MFCC1 emerging as the most discriminative features. The findings highlight the efficacy of ensemble methods in capturing nuanced emotional cues in Indonesian speech, outperforming prior studies on locally sourced datasets. Practical implications include applications in customer service analytics and mental health tools, though limitations such as the dataset’s-controlled conditions and fixed sentence structure necessitate caution in real-world deployment. Future work should expand the dataset to include regional dialects, spontaneous speech, and hybrid architectures like CNN-LSTMs. This study establishes foundational benchmarks for Indonesian SER, advocating for culturally informed models to enhance human-computer interaction in underrepresented linguistic contexts.