Article Per Year (5 Year)
p-Index From 2021 - 2026
0.817
P-Index
This Author published in this journals
All Journal Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) KOMPUTIKA - Jurnal Sistem Komputer Djtechno: Jurnal Teknologi Informasi International Journal of Social Science, Educational, Economics, Agriculture Research, and Technology (IJSET)
Muhamad Ridwan
Program Studi Teknik Informatika, Fakultas Sains Dan Teknologi, Universitas Muhammadiyah Kalimantan Timur
Agriculture, Biological Sciences & Forestry Humanities Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Education Electrical & Electronics Engineering Engineering

Published : 4 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 4 Documents
Search

 1 
An Optimized Hyperparameter Tuning for Improved Hate Speech Detection with Multilayer Perceptron Muhamad Ridwan; Ema Utami
Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi) Vol 8 No 4 (2024): August 2024
Publisher : Ikatan Ahli Informatika Indonesia (IAII)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29207/resti.v8i4.5949

Abstract

Hate speech classification is a critical task in the domain of natural language processing, aiming to mitigate the negative impacts of harmful content on digital platforms. This study explores the application of a Multilayer Perceptron (MLP) model for hate speech classification, utilizing Bag of Words (BoW) for feature extraction. The hypothesis posits that hyperparameter tuning through sophisticated optimization techniques will significantly improve model performance. To validate this hypothesis, we employed two distinct hyperparameter tuning approaches: Random Search and Optuna. Random Search provides a straightforward yet effective means of exploring the hyperparameter space, while Optuna offers a more sophisticated, optimization-based approach to hyperparameter selection. The study involved training the MLP model on a labeled dataset is based on crawling results on the Twitter platform of hate speech and non-hate speech overall total dataset is 13.169, followed by evaluation using standard metrics. Our experimental results demonstrate the comparative effectiveness of these two hyperparameter tuning methods. Notably, the MLP model tuned with Optuna achieved a higher F1-score of 81.49%, compared to 79.70% achieved with Random Search, indicating the superior performance of Optuna in optimizing the hyperparameters. These results were obtained through extensive cross-validation to ensure robustness and generalizability. The findings underscore the importance of optimized hyperparameters in developing robust hate speech classification systems. The superior perform ance of Optuna highlights its potential for broader application in other machine learning tasks requiring hyperparameter optimization. This improvement enables more reliable and efficient automated moderation, which is crucial for the integrity and security of digital communication platforms such as Twitter.
Optimasi Prediksi Kelayakan Pinjaman dengan Teknik Resampling dan Algoritma Boosting Putra, Muhammad Ricky Perdana; Juwariyah, Siti; Ridwan, Muhammad; Marco, Robert
Komputika : Jurnal Sistem Komputer Vol. 14 No. 2 (2025): Komputika: Jurnal Sistem Komputer
Publisher : Computer Engineering Departement, Universitas Komputer Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.34010/komputika.v14i2.15485

Abstract

Loan eligibility assessment is a crucial element in financial risk mitigation, aiming to minimize potential losses due to bad debts and ensure proper resource distribution. Traditional rule-based approaches have limitations in scalability, risk of subjective bias, and complex data management. The application of Machine Learning (ML) presents a solution with the ability to analyze complex patterns in historical data, although significant challenges such as class imbalance where the number of defaulted borrowers is much smaller than that of current borrowers and missing values ​​in the dataset remain major obstacles. This study evaluates the SMOTE and SMOTE-ENN resampling methods, to address class imbalance, as well as the mean imputation technique to handle missing values. By evaluating boosting algorithms, including Gradient Boosting, XGBoost, LightGBM, AdaBoost, and CatBoost, the results show that the combination of the CatBoost algorithm with the SMOTE-ENN sampling technique provides the highest prediction accuracy of 91.67%. This finding confirms the significant potential of ML in improving the accuracy, efficiency, and fairness of predictions, while making important contributions to the development of data-driven decision-making systems in the financial sector.
IMPLEMENTATION OF GEMINI PRE-PROCESSING ON 2024 SIREKAP REVIEWS USING THE RANDOM FOREST ALGORITHM Amru Omar; Naufal Azmi Verdikha; Muhamad Ridwan
International Journal of Social Science, Educational, Economics, Agriculture Research and Technology (IJSET) Vol. 5 No. 3 (2026): FEBRUARY
Publisher : RADJA PUBLIKA

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.5281/zenodo.18939969

Abstract

This study aims to classify reviews of the SIREKAP 2024 application by utilizing Large Language Model (LLM)-based Gemini pre-processing, Term Frequency–Inverse Document Frequency (TF-IDF) feature extraction, and the Random Forest algorithm as the classification method. The data used consist of user reviews obtained from the Google Play Store and categorized into five rating classes. Model performance evaluation was conducted using the 10-Fold Cross-Validation method with the Macro F1-Score metric. The testing results indicate that the lowest F1-Score achieved was 31.87%, while the highest reached 37.28%, with an overall average Macro F1-Score of 34.62%. These findings demonstrate that the Random Forest algorithm is capable of producing relatively stable classification performance through its ensemble learning mechanism, which combines multiple decision trees. However, its performance is still influenced by the imbalance in data distribution across classes. Therefore, Random Forest plays a role in maintaining prediction stability and reducing overfitting, although further development is required to improve classification performance on imbalanced review data
EKSPLORASI PRA-PROSES LARGE LANGUAGE MODEL GEMINI 2.0 UNTUK KLASIFIKASI ULASAN SIREKAP MENGGUNAKAN SVM Muhamad Ridwan
Djtechno: Jurnal Teknologi Informasi Vol 7, No 2 (2026): Agustus
Publisher : Universitas Dharmawangsa

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.46576/djtechno.v7i2.9714

Abstract

Implementasi aplikasi SIREKAP 2024 dalam Pemilihan Umum menuai beragam respon publik. Ulasan Google Play Store menjadi data krusial untuk mengevaluasi kepuasan pengguna melalui rating. Studi ini bertujuan mengklasifikasikan ulasan SIREKAP ke dalam lima kategori rating. Metode yang digunakan adalah Support Vector Machine (SVM) dengan pembobotan TF-IDF. Kebaruan penelitian ini terletak pada pemanfaatan API Gemini 2.0 Flash (LLM) untuk tahap pra-proses, yang menangani koreksi ejaan dan normalisasi kata secara kontekstual. Analisis menunjukkan ketidakseimbangan kelas yang signifikan, didominasi oleh rating 1 sebesar 64,14%. Berdasarkan pengujian K-Fold Cross Validation, model LinearSVC menghasilkan rata-rata F1-Score sebesar 36,68%. Skor ini menunjukkan kesulitan model dalam memprediksi kelas, sehingga memerlukan teknik penanganan lanjutan dalam penelitian mendatang untuk meningkatkan performa.
Co-Authors Amru Omar Ema Utami Juwariyah, Siti Naufal Azmi Verdikha Putra, Muhammad Ricky Perdana Robert Marco, Robert