Article Per Year (5 Year)
p-Index From 2021 - 2026
1.513
P-Index
This Author published in this journals
All Journal JOURNAL OF SCIENCE AND SOCIAL RESEARCH Jurnal Teknik Informatika (JUTIF) Jurnal Minfo Polgan (JMP) Journal of Technology and Computer (JOTECHCOM) LogicLink: Journal of Artificial Intelligence and Multimedia in Informatics International Journal of Educational Insights and Innovations (IJEDINS)
Manza, Yuke
Unknown Affiliation
Computer Science & IT Economics, Econometrics & Finance Education Languange, Linguistic, Communication & Media Library & Information Science Mathematics Social Sciences Other

Published : 9 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 9 Documents
Search

 1 
AI for MSMEs: Smart Solutions to Optimize Operations and Marketing Manza, Yuke
Journal of Technology and Computer Vol. 2 No. 2 (2025): May 2025 - Journal of Technology and Computer
Publisher : PT. Technology Laboratories Indonesia (TechnoLabs)

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

Abstract

This research investigates the transformative potential of Artificial Intelligence (AI) for Micro, Small, and Medium Enterprises (MSMEs) in optimizing operations and marketing. Employing a mixed-methods approach—combining quantitative surveys (n=100+) and in-depth qualitative interviews (n=15-20)—the study reveals a significant positive correlation between AI adoption and enhanced operational efficiency, evidenced by average reductions of 25% in data processing time and 15% in inventory management. Furthermore, AI substantially boosts marketing effectiveness, leading to a 30% increase in audience reach and an 18% rise in sales conversion rates. Despite these clear benefits, MSMEs face considerable barriers to AI adoption, primarily financial constraints (65% of respondents) and limited digital literacy (58%). To address these challenges, the research proposes an affordable and easy-to-implement AI framework emphasizing cloud-based solutions (SaaS) and comprehensive training programs. The findings underscore AI as a crucial driver for MSME competitiveness and recommend concerted efforts from government and industry stakeholders to foster a supportive ecosystem. This study bridges the digital divide, offering evidence-based recommendations for resilient, efficient, and sustainable MSMEs in the digital era.
Expert System untuk Rekomendasi Pemilihan Bahasa Pemrograman bagi Pemula Menggunakan Algoritma Decision Tree Manza, Yuke; Wayahdi, M. Rhifky
LogicLink Vol. 2 No. 1, Juni 2025
Publisher : Universitas Islam Negeri K.H. Abdurrahman Wahid Pekalongan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.28918/logiclink.v2i1.10924

Abstract

This study develops an expert system based on the Decision Tree algorithm to recommend suitable programming languages for beginners, addressing the challenge of selecting the right language amid the abundance of options and diverse learning goals. This topic is significant because choosing the appropriate language can accelerate the learning process and improve the effectiveness of programming education. The research methodology includes the creation of a synthetic dataset comprising 1,500 entries, with the addition of 5% noise. This noise is introduced to simulate real-world data imperfections and to test the model's robustness against unclean or imperfect data. The next stages involve data preprocessing through encoding and normalization, followed by modeling using the Decision Tree algorithm with hyperparameter optimization to enhance model performance. Evaluation results show an accuracy of 95%, with learning goals (38% contribution) and platform preference (35%) emerging as the most influential factors in decision-making. A 10-fold cross-validation produced an average error of 0.046, indicating model stability across various data subsets. Feature importance analysis revealed that the model logically prioritizes technical relevance, for example, by ranking learning goals and platform preference above demographic features, as these are more directly related to the context and practical use of programming languages. The implemented system successfully provided relevant recommendations, such as Python for Data Science and JavaScript for Web Development. This study concludes that the Decision Tree algorithm is effective for recommendation systems based on user profiles, although data enhancement is needed for minority classes such as Java. These findings contribute to the development of more personalized and adaptive programming learning support tools.
Text Classification Using TF-IDF and Naïve Bayes: Case Study of MyXL App User Review Data Nurhayati, Nurhayati; Hartimar, Lima; Manza, Yuke; Siregar, Kiki Putriani
Journal of Technology and Computer Vol. 2 No. 2 (2025): May 2025 - Journal of Technology and Computer
Publisher : PT. Technology Laboratories Indonesia (TechnoLabs)

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

Abstract

The MyXL application, developed by leading Indonesian operator XL Axiata, allows customers to independently manage their telecommunication services. However, a significant volume of negative user reviews necessitates a deeper analysis of user sentiment. This research classifies MyXL app reviews using the TF-IDF (Term Frequency-Inverse Document Frequency) method for feature extraction and the Naïve Bayes algorithm for sentiment classification, implemented via a Python-based GUI. The study's objective is to categorize reviews into positive, negative, and neutral sentiments. A dataset of 1000 user reviews from Kaggle underwent comprehensive preprocessing—including text cleaning, normalization, tokenization, stopword removal, and stemming—before conversion into a numerical representation using TF-IDF. The classification model, built with the Naïve Bayes algorithm, was evaluated using accuracy, precision, recall, and F1-score metrics. The model achieved an accuracy of 61.5%. This finding demonstrates that combining TF-IDF and Naïve Bayes is effective for classifying sentiment in Indonesian text reviews, particularly within the mobile app domain. Furthermore, the methodology shows clear potential for development into a large-scale and automated user opinion analysis system.
EVALUASI DENSENET-201 UNTUK IDENTIFIKASI BIJI KOPI MENGGUNAKAN HYPERPARAMETER GRIDSEARCH Manza, Yuke; Rambe, Lima Hartimar; Siregar, Kiki Putri Ani; Rosnelly, Rika; Setiawan, Adil
JOURNAL OF SCIENCE AND SOCIAL RESEARCH Vol 8, No 3 (2025): August 2025
Publisher : Smart Education

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54314/jssr.v8i3.3898

Abstract

Abstract: Coffee is one of the most important commodities in the global agricultural sector. However, the manual sorting process of coffee beans, which is still widely applied in the Small and Medium Industry (IKM) sector, tends to be time-consuming and often results in inconsistent quality assessments. This study aims to classify coffee bean quality using the DenseNet-201 deep learning architecture, optimized with the GridSearch method to obtain the best combination of hyperparameters. The dataset used consists of 450 images of coffee beans divided into two classes: good-quality and defective beans. The model was trained for 20 epochs using a transfer learning approach and evaluated using performance metrics such as accuracy, precision, recall, and F1-score. The test results show that the model before optimization achieved an accuracy of only 78.67%, while the model optimized with GridSearch reached a high accuracy of 99.47% with a low loss value. These findings indicate that the application of DenseNet-201 with hyperparameter tuning is capable of producing accurate and stable classification results, and can be relied upon as an automated solution for sorting coffee beans based on their quality. Keywords: Deep Learning, DenseNet201, Hyperparameter, GridSearch, Coffee Bean Classification Abstrak: Kopi merupakan salah satu komoditas penting dalam sektor pertanian global. Namun, proses pemilahan biji kopi secara manual yang masih banyak diterapkan pada sektor Industri Kecil dan Menengah (IKM) cenderung memakan waktu dan menghasilkan penilaian kualitas yang tidak konsisten. Penelitian ini bertujuan untuk mengklasifikasikan kualitas biji kopi menggunakan arsitektur Deep Learning DenseNet-201 yang dioptimalkan dengan metode GridSearch untuk memperoleh kombinasi hyperparameter terbaik. Dataset yang digunakan terdiri dari 450 gambar biji kopi dengan dua kelas: biji kopi bagus dan biji kopi rusak. Model dilatih selama 20 epoch dengan pendekatan transfer learning dan dilakukan evaluasi terhadap performa model menggunakan metrik akurasi, precision, recall, dan f1-score. Hasil pengujian menunjukkan bahwa model sebelum optimasi hanya mencapai akurasi sebesar 78,67%, sedangkan model dengan optimasi GridSearch mampu mencapai akurasi tinggi sebesar 99,47% dan nilai loss yang rendah. Hal ini menunjukkan bahwa penerapan DenseNet-201 dengan tuning hyperparameter mampu menghasilkan klasifikasi yang akurat dan stabil, serta dapat diandalkan sebagai solusi otomatis dalam proses sortasi biji kopi berdasarkan kualitasnya. Kata kunci: Deep Learning, DenseNet201, Hyperparameter, GridSearch, Klasifikasi Biji Kopi
KLASIFIKASI BIJI KOPI MENGGUNAKAN TEKNIK KOMBINASI RANDOM FOREST DAN INCEPTION V3 UNTUK EKSTRAKSI FITUR Rambe, Lima Hartimar; Manza, Yuke; Ashari, Annisa; Negoro, Wahyu Saptha
JOURNAL OF SCIENCE AND SOCIAL RESEARCH Vol 8, No 3 (2025): August 2025
Publisher : Smart Education

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54314/jssr.v8i3.3976

Abstract

Abstract: Coffee bean classification is a crucial step in ensuring the quality and selling value of coffee products. Manual sorting methods are often inefficient and error-prone, necessitating a technology-based automated approach. This study proposes a combination of the Inception V3 architecture as an image feature extraction method and the random forest algorithm as a classifier to distinguish good and defective coffee beans. The dataset used consists of 986 images, divided into training and test data. The processing was carried out using the Orange Data Mining platform, which includes pre-processing, feature extraction, model training, and performance evaluation. The evaluation results show that the model produces an accuracy of 96.4% on the training data and 96.8% on the test data. In addition, other performance metrics such as AUC (1.000), F1-score (0.967), precision (0.968), recall (0.968), and MCC (0.922) strengthen the model's excellent classification performance. Thus, the combined approach of Inception V3 and random forest is proven effective and has the potential to be implemented in a digital image-based coffee bean classification system. Keywords: Coffee Bean Classification, Random Forest, Inception V3, Feature Extraction, Digital Imagery Abstrak: Klasifikasi biji kopi merupakan langkah penting dalam menjamin mutu dan nilai jual produk kopi. Metode manual dalam penyortiran sering kali tidak efisien dan rentan kesalahan, sehingga dibutuhkan pendekatan otomatis berbasis teknologi. Penelitian ini mengusulkan kombinasi arsitektur Inception V3 sebagai metode ekstraksi fitur citra dan algoritma random forest sebagai klasifikator untuk membedakan biji kopi bagus dan rusak. Dataset yang digunakan terdiri dari 986 gambar, terbagi menjadi data latih dan data uji. Proses pengolahan dilakukan menggunakan platform Orange Data Mining, yang meliputi tahap pra-pemrosesan, ekstraksi fitur, pelatihan model, dan evaluasi kinerja. Hasil evaluasi menunjukkan bahwa model menghasilkan akurasi sebesar 96,4% pada data latih dan 96,8% pada data uji. Selain itu, metrik performa lain seperti AUC (1.000), F1-score (0.967), precision (0.968), recall (0.968), dan MCC (0.922) memperkuat bahwa model ini memiliki kinerja klasifikasi yang sangat baik. Dengan demikian, pendekatan kombinasi Inception V3 dan random forest terbukti efektif dan berpotensi diimplementasikan dalam sistem klasifikasi biji kopi berbasis citra digital. Kata kunci: Klasifikasi Biji Kopi, Random Forest, Inception V3, Ekstraksi Fitur, Citra Digital
Model Machine Learning untuk Klasifikasi Warna Fashion Menggunakan Metode K-Nearest Neighbor Manza, Yuke; Suhada WD, Muhammad; Ndruru, Agus F.S.; Rosnelly, Rika
Jurnal Minfo Polgan Vol. 13 No. 2 (2024): Artikel Penelitian
Publisher : Politeknik Ganesha Medan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33395/jmp.v13i2.14551

Abstract

Penelitian ini bertujuan untuk mengembangkan model klasifikasi warna fashion menggunakan metode K-Nearest Neighbor (KNN). Dengan mengumpulkan data dari 25.000 sampel yang mencakup variabel tinggi badan, berat badan, jenis kelamin, indeks massa tubuh, dan warna kulit, penelitian ini menerapkan tahapan metodologi yang sistematis, termasuk pengumpulan data, pra-pemrosesan, pemodelan, dan evaluasi. Hasil analisis menunjukkan bahwa model KNN mencapai akurasi sebesar 78.22%, dengan kemampuan yang baik dalam membedakan kelas positif dan negatif. Meskipun hasil ini memuaskan, terdapat peluang untuk meningkatkan akurasi model melalui optimasi parameter atau penggabungan dengan algoritma lain. Temuan ini memberikan kontribusi signifikan terhadap pengembangan teknologi machine learning dalam industri fashion, serta menawarkan solusi praktis bagi desainer dan produsen dalam memilih kombinasi warna yang menarik, sehingga dapat meningkatkan daya tarik produk di pasar.
The Impact of Active Learning Methods on Student Motivation and Academic Achievement in Elementary Schools Yusof, Khairunnisa M.; Manza, Yuke
International Journal of Educational Insights and Innovations Vol. 1 No. 2 (2024): December 2024 - International Journal of Educational Insights and Innovations (
Publisher : PT. Technology Laboratories Indonesia (TechnoLabs)

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

Abstract

This study investigates the impact of active learning methods on student motivation and academic achievement in elementary schools. With a shift from traditional instructional approaches to more interactive and student-centered learning, this research employs a mixed-methods design, combining quantitative surveys and qualitative interviews. A structured survey was administered to 200 elementary students to assess their motivation levels and academic performance before and after the implementation of active learning strategies. Additionally, semi-structured interviews were conducted with 15 teachers to gain insights into their experiences with these methods. The findings reveal a significant increase in student motivation, with 75% of participants reporting enhanced engagement, alongside a 15% improvement in average test scores post-intervention. Qualitative data indicate that active learning fosters collaboration and communication skills among students, contributing to a richer classroom environment. Despite challenges related to training and resource availability, educators overwhelmingly support the integration of active learning techniques. This research highlights the effectiveness of active learning in promoting not only academic success but also essential soft skills, thereby advocating for its broader application in educational settings to create more engaging and productive learning experiences for young learners.
PREDIKSI PENJUALAN SUPERMARKET MENGGUNAKAN JARINGAN SYARAF TIRUAN LONG SHORT-TERM MEMORY (LSTM) Rambe, Lima Hartimar; Manza, Yuke; Siregar, Kiki Putri Ani; Roslina, Roslina
JOURNAL OF SCIENCE AND SOCIAL RESEARCH Vol 9, No 1 (2026): February 2026
Publisher : Smart Education

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.54314/jssr.v9i1.5807

Abstract

Abstract: Sales forecasting is a crucial aspect of supermarket operations, as it supports inventory management, production planning, and strategic decision-making. Sales data typically exhibit complex patterns such as trends, seasonality, and fluctuations, requiring modeling methods capable of handling nonlinear time-series characteristics. This study employs the Long Short-Term Memory (LSTM) model, an advanced form of Recurrent Neural Network (RNN) designed to capture long-term dependencies and overcome the vanishing gradient problem. The secondary dataset was obtained from the Kaggle platform, consisting of 20 features and a total of 1,000 records. The LSTM model was constructed using 50 neurons in the LSTM layer and a single dense output layer. Training was conducted for 100 epochs using the Adam optimizer and Mean Squared Error (MSE) as the loss function. The training process showed a consistent decrease in loss, reaching approximately 0.0193, while evaluation using Root Mean Squared Error (RMSE) indicated that the model effectively learned historical patterns. Visualization of predictions on the test dataset demonstrated that the model successfully followed sales trends, although it was less responsive to extreme fluctuations. Overall, the LSTM model proved effective for daily sales forecasting and can serve as a valuable tool for operational planning in supermarkets. Keywords: LSTM, Sales Forecasting, Time Series, Deep Learning, RMSE, Supermarket. Abstrak: Peramalan penjualan merupakan aspek penting dalam operasional supermarket karena berperan besar dalam pengelolaan inventaris, perencanaan produksi, serta pengambilan keputusan strategis. Data penjualan umumnya memiliki pola tren, musiman, dan fluktuasi yang kompleks sehingga memerlukan metode pemodelan yang mampu menangani karakteristik deret waktu nonlinear. Penelitian ini menggunakan model Long Short-Term Memory (LSTM), sebuah pengembangan Recurrent Neural Network (RNN) yang efektif dalam menangkap dependensi jangka panjang dan mengatasi masalah vanishing gradient. Data sekunder diperoleh dari platform Kaggle dengan 20 fitur dan total 1.000 record. Model LSTM dibangun menggunakan 50 unit neuron pada lapisan LSTM dan satu lapisan dense sebagai output. Model dilatih selama 100 epoch menggunakan optimizer Adam dan fungsi loss MSE. Hasil pelatihan menunjukkan penurunan loss yang stabil hingga mencapai nilai sekitar 0,0193, sedangkan evaluasi menggunakan Root Mean Squared Error (RMSE) menunjukkan bahwa model mampu mempelajari pola historis dengan baik. Visualisasi prediksi pada data pengujian memperlihatkan bahwa model mampu mengikuti tren pergerakan penjualan meskipun masih kurang responsif terhadap fluktuasi ekstrem. Secara keseluruhan, model LSTM terbukti efektif dalam memprediksi penjualan harian dan dapat digunakan sebagai dasar pengambilan keputusan dalam perencanaan operasional supermarket. Kata kunci: LSTM, Peramalan Penjualan, Deret Waktu, Deep Learning, RMSE, Supermarket.
Optimized KNN Performance with PCA and K-Fold Cross-Validation for Colorectal Cancer Survival Prediction Manza, Yuke; Rosnelly, Rika; Furqan, Mhd; Riza, Bob Subhan
Jurnal Teknik Informatika (Jutif) Vol. 7 No. 1 (2026): JUTIF Volume 7, Number 1, February 2026
Publisher : Informatika, Universitas Jenderal Soedirman

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52436/1.jutif.2026.7.1.5422

Abstract

Colorectal cancer remains a leading cause of global mortality, necessitating effective predictive tools for patient survival. While Machine Learning algorithms like K-Nearest Neighbors (KNN) utilize patient data for prediction, standard KNN implementations often suffer from the curse of dimensionality and overfitting, leading to unreliable performance on complex medical datasets. This study aims to evaluate and optimize the performance of the KNN algorithm by integrating Principal Component Analysis (PCA) for dimensionality reduction and K-Fold Cross-Validation (KFCV) to enhance model stability. The research utilized a quantitative approach on a global colorectal cancer dataset, processing demographic and clinical features through a rigorous pipeline of imputation, encoding, and normalization. Three model configurations were systematically compared: Standard KNN, KNN combined with PCA, and an optimized KNN model utilizing both PCA and KFCV across various neighbor values. The results demonstrate a distinct trade-off between predictive sensitivity and model stability. While the Standard KNN and PCA-enhanced models achieved higher recall, indicating a strong ability to identify survivors in a single data split, the fully optimized KNN+PCA+KFCV model provided the most stable and generalized accuracy with minimal deviation. These findings indicate that while PCA effectively reduces computational complexity without information loss, the integration of cross-validation is crucial for obtaining an honest assessment of model performance. This research contributes to clinical informatics by highlighting the necessity of prioritization between high sensitivity and generalization stability when developing survival prediction models for complex, inseparable medical data.
Co-Authors Ashari, Annisa Bob Subhan Riza, Bob Subhan Hartimar, Lima M. Rhifky Wayahdi Mhd Furqan Ndruru, Agus F.S. Negoro, Wahyu Saptha Nurhayati Nurhayati Rambe, Lima Hartimar Rika Rosnelly, Rika Roslina Roslina, Roslina Setiawan, Adil Siregar, Kiki Putri Ani Siregar, Kiki Putriani Suhada WD, Muhammad Yusof, Khairunnisa M.