Article Per Year (5 Year)
p-Index From 2021 - 2026
1.048
P-Index
This Author published in this journals
All Journal Journal Computer Science and Informatic Systems : J-Cosys Jurnal Ilmiah Betrik : Besemah Teknologi Informasi dan Komputer Media Journal of General Computer Science (MJGCS) CODEX: Journal of Software Engineering
Yusi Nurmala Sari
Universitas Serelo Lahat
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Electrical & Electronics Engineering

Published : 6 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 6 Documents
Search

 1 
Optimasi Hyperparameter WOA-SVM pada Citra Daun Kopi Terpupuk NPK Agustian Prakarsya; Nina Dwi Putriani; Yusi Nurmala Sari; Firza Septian
BETRIK Vol. 16 No. 02 (2025): Jurnal Ilmiah BETRIK : Besemah Teknologi Informasi dan Komputer
Publisher : PPPM Institut Teknologi Pagar Alam

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36050/zrj1e094

Abstract

This study aims to analyze the impact of Whale Optimization Algorithm (WOA) optimization on the performance of Support Vector Machine (SVM) in classifying images of coffee leaves treated with NPK fertilizer. WOA is employed to find the optimal combination of SVM parameters to improve classification accuracy. The dataset consists of coffee leaf images that have undergone feature extraction based on color and texture. Performance evaluation was conducted using a confusion matrix, classification report, and heatmap visualization. The results show that the SVM model optimized with WOA performs better than the non-optimized SVM. Specifically, the non-optimized SVM achieved a precision of 0.82, recall of 0.81, and F1-score of 0.81. After optimization with WOA, the model’s precision increased to 0.90, recall to 0.88, and F1-score to 0.87. This study demonstrates that metaheuristic approaches like WOA can significantly enhance the performance of classification algorithms in the context of digital image processing. The findings have practical implications for early detection of plant quality through image-based analysis in technology-driven agriculture
A Rule-Based AI Writing Assistant for Beginner English Learners with Visual Feedback Arief Zikry; Yusi Nurmala Sari; Muhammad Sulkhan Nurfatih; Firza Septian
Media Journal of General Computer Science Vol. 3 No. 1 (2026): MJGCS
Publisher : MASE - Media Applied and Engineering

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62205/mjgcs.v3i1.149

Abstract

The increasing adoption of artificial intelligence (AI) in educational technology has created new opportunities to support second language (L2) writing development. Beginner English learners often struggle with grammatical accuracy, limited vocabulary, and unclear sentence construction, while immediate and individualized feedback remains difficult to provide in traditional learning settings. This study proposes a rule-based AI writing assistant designed to deliver automated, transparent, and interpretable feedback for beginner-level English writing without relying on data-intensive machine learning models. The system employs symbolic AI principles through predefined grammatical rules and heuristic textual metrics to evaluate writing quality across three dimensions: grammar accuracy, vocabulary richness, and text clarity. Grammar errors are detected using regular expression-based rules, vocabulary quality is measured via lexical diversity ratios, and clarity is estimated using a length-based heuristic. These metrics are normalized and combined to produce an overall writing quality score. To enhance usability and learner engagement, the system integrates visual feedback elements, including progress bars, graphical score representations, and animated character responses. Functional testing using sample beginner texts demonstrates that the proposed system effectively identifies common writing issues, provides consistent scoring, and delivers immediate, explainable feedback. The results indicate that rule-based AI, when combined with visual feedback mechanisms, can offer a lightweight, efficient, and pedagogically meaningful solution for beginner English writing support. This approach is particularly suitable for educational contexts that prioritize explainability, accessibility, and low computational requirements.
Predicting Purchase Decision Using a Hybrid KNN-WOA Model Based on Social Media Marketing and Word of Mouth Quality Yusi Nurmala Sari; Nina Dwi Putriani; Agustian Prakarsya; Firza Septian
Journal Computer Science and Information Systems : J-Cosys Vol 5, No 2 (2025): September
Publisher : Universitas Dharma Wacana

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.53514/jco.v5i2.701

Abstract

Penelitian ini bertujuan untuk mengembangkan dan menguji model hybrid K-Nearest Neighbor (KNN) yang dioptimasi dengan Whale Optimization Algorithm (WOA) dalam memprediksi keputusan pembelian konsumen berdasarkan variabel Social Media Marketing (SMM) dan Word of Mouth Quality (WQ). Data penelitian diperoleh dari 100 responden dengan 22 indikator yang diukur menggunakan skala Likert 1–7. Variabel dependen berupa Purchase Decision dibentuk dari lima indikator dan dikonversi menjadi kelas biner untuk keperluan klasifikasi. Hasil analisis deskriptif menunjukkan bahwa indikator SMM dan WQ memiliki distribusi yang stabil dengan kecenderungan nilai tinggi, serta korelasi positif terhadap keputusan pembelian. Model hybrid KNN–WOA menghasilkan akurasi sebesar 95% dengan precision 0.95, recall 1.00, dan f1-score 0.97 pada kelas positif. Temuan ini menegaskan bahwa kualitas konten media sosial dan kredibilitas informasi Word of Mouth berperan signifikan dalam memengaruhi keputusan pembelian konsumen. Penelitian ini memberikan kontribusi teoritis dalam pengembangan model prediktif berbasis optimasi metaheuristik serta kontribusi praktis bagi perusahaan dalam merancang strategi pemasaran digital yang lebih efektif dan berbasis data.
Explainable Boosted Ensemble Penjualan Video Game Release Tahun 1980-2020 Nina Dwi Putriani; Yusi Nurmala Sari; Selvy Megira
BETRIK Vol. 17 No. 02 (2026): Jurnal Ilmiah BETRIK : Besemah Teknologi Informasi dan Komputer
Publisher : PPPM Institut Teknologi Pagar Alam

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36050/8a771z88

Abstract

This study aims to analyze video game sales trends of game released on 1980 to 2020 using the Explainable Boosted Ensemble approach. The XGBoost algorithm was selected for its strong predictive ability on tabular data, while SHAP integration provides transparency regarding the factors influencing predictions. The dataset includes variables such as genre, platform, publisher, and both regional and global sales, enabling a comprehensive analysis of market preferences in North America, Europe, Japan, and other regions. Findings reveal that regional sales, particularly in North America and Europe, contribute most significantly to global sales, while Japan shows dominance in Role-Playing and Platform genres. Model evaluation produced an R² score of 0.7788, indicating reliable accuracy in explaining sales variations. Furthermore, genre recommendations highlight Platform, Shooter, and Role-Playing as the backbone of the industry, with Action, Racing, Fighting, and Sports remaining relevant in specific segments. This research is expected to provide both academic and practical contributions, offering insights for developers to design more effective distribution strategies and genre portfolios.
Prediksi Risiko Penyakit Jantung dengan Decision Tree yang Dioptimasi Algoritma Bald Eagle Search Yusi Nurmala Sari; Selvy Megira; Salamudin Salamudin
BETRIK Vol. 17 No. 02 (2026): Jurnal Ilmiah BETRIK : Besemah Teknologi Informasi dan Komputer
Publisher : PPPM Institut Teknologi Pagar Alam

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36050/d0q2c524

Abstract

Heart disease remains one of the leading causes of death worldwide, making early detection of its risk crucial to reducing mortality and morbidity rates. This study aims to develop a heart disease risk prediction model based on machine learning using a Decision Tree algorithm optimized with Bald Eagle Search (BES). The research employed a quantitative approach utilizing a clinical dataset containing demographic and medical variables such as age, gender, blood pressure, cholesterol levels, electrocardiographic results, and heart disease status. The baseline Decision Tree model was compared with the BES-optimized model (BES-DT) through evaluations of accuracy, confusion matrix, prediction probability distribution, feature importance analysis, and learning curves with respect to the max_depth parameter. The analysis revealed that the baseline Decision Tree achieved an accuracy of 70.5%, with 43 correct predictions out of 61 test samples, while the BES-DT model achieved an accuracy of 68.9%, with 42 correct predictions. Although the overall accuracy showed a slight decrease, BES-DT demonstrated greater consistency in identifying at-risk patients, with fewer misclassifications (4 cases compared to 7 in the baseline). Furthermore, the prediction probability distribution in BES-DT was more stable, with values concentrated near 0 and 1, indicating higher confidence in classification. The feature importance analysis highlighted chest pain type, oldpeak, and thal as dominant variables in risk classification. The learning curve confirmed that BES-DT reduced the risk of overfitting and improved the model’s generalization capability. This study contributes to the development of more accurate and interpretable machine learning classification methods in healthcare. Future work may involve testing the model on larger and more diverse datasets, integrating other optimization algorithms for performance comparison, and implementing web-based or clinical applications to support medical decision-making.
BES-OPTIMIZED KNN FOR PURCHASE DECISION PREDICTION BASED ON SOCIAL MEDIA MARKETING AND WORD OF MOUTH QUALITY Firza Septian; Yusi Nurmala Sari; Nina Dwi Putriani
CODEX: Journal of Software Engineering Vol 1, No 2 (2026): June 2026
Publisher : Nurul Jadid University, Paiton Probolinggo, East Java

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33650/codex.v1i2.15767

Abstract

This study proposes a Hybrid Machine Learning Framework for purchase decision analysis by integrating the Bald Eagle Search (BES) algorithm with the K-Nearest Neighbors (KNN) classifier. The dataset consists of 500 respondents collected through survey instruments, with Social Media Marketing (SMM) and Word of Mouth Quality (WQ) serving as predictor features, while Purchase Perception (PP) items are aggregated to form the target variable. The rapid growth of digital marketing and social media has made indicators such as SMM, which reflects consumer engagement with online marketing activities, and WQ, which captures the credibility of peer recommendations, critical in influencing consumer trust and purchase behavior. To capture these nonlinear relationships, the framework applies systematic data preparation, preprocessing with SMOTE balancing and standardization, and BES optimization to determine the optimal neighborhood size K for KNN. The dataset was split into training and testing subsets with an 80:20 ratio, resulting in 50 test samples used for evaluation each WOA-KNN and BES-KNN. The evaluation was conducted using accuracy, precision, recall, F1-score, and ROC curve analysis. Results show that the hybrid BES-KNN model achieved 96% accuracy, with precision, recall, and F1-score all at approximately 97.9%, indicating balanced and reliable classification performance; however, the moderate AUC value of 0.68 suggests that class imbalance and the distribution of prediction probabilities may have influenced discriminatory power, which explains the apparent inconsistency between high classification metrics and the ROC curve. This framework contributes to advancing machine learning applications in marketing analytics and provides a foundation for future research to refine feature selection and enhance discriminatory power
Co-Authors Agustian Prakarsya Firza Septian Muhammad Sulkhan Nurfatih Nina Dwi Putriani salamudin salamudin Selvy megira