Claim Missing Document
Check
Articles

Comparative Study of Time Series Forecasting on Iron Sales Using CNN, MLP, and LSTM Nabila Putri Listyanto; Yustanti, Wiyli
Journal of Emerging Information Systems and Business Intelligence (JEISBI) Vol. 6 No. 3 (2025): Vol. 06 Issue 03
Publisher : Universitas Negeri Surabaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26740/jeisbi.v6i3.71361

Abstract

Sales forecasting is essential for businesses to predict future demand and inform strategic and operational planning, especially in the building materials retail industry. Accurate sales prediction supports inventory management, cost control, and supply chain efficiency. This study compares the performance of 3 deep learning models, Convolutional Neural Network (CNN), Multilayer Perceptron (MLP), and Long Short-Term Memory (LSTM), in forecasting daily iron sales at PT Surya Aneka Bangunan from 2016 to 2020. The models were trained on 80% of the historical data and tested on 20%. Model performance was evaluated using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and Coefficient of Determination R². The results show that the CNN model achieved the best performance with an MAE of 0.293, RMSE of 0.357, MAPE of 0.081, and R² of 0.9989, indicating high accuracy and stability. The MLP model produced higher errors, while the LSTM model had the lowest MAPE but greater error variability. These findings suggest that the CNN model is the most reliable for capturing temporal patterns in iron sales data. The study contributes to the development of adaptive sales forecasting systems and opens opportunities for applying similar methods in other retail sectors to support data driven decision making.
Detection of Dirty Bowel Disease Through Palm Image Analysis Using CNN-VGG16 Algorithm Kurniasari, Calycha; Yustanti, Wiyli
Journal of Emerging Information Systems and Business Intelligence (JEISBI) Vol. 6 No. 3 (2025): Vol. 06 Issue 03
Publisher : Universitas Negeri Surabaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26740/jeisbi.v6i3.71712

Abstract

Early detection of disease is very important in improving the quality of human health. The quality of life of patients suffering from gross bowel disease can be significantly affected, including daily activities, work, and interpersonal relationships. One promising innovative method in the healthcare field is disease detection through palm image analysis. The solution to this problem is done by implementing the Convolutional Neural Network (CNN) algorithm using the VGG16 architecture model which can be operated by uploading palm images to detect Dirty Bowel Disease, Other Diseases (Not Dirty Bowel), and Healthy Hands through a web-based application. Based on the test results, the test accuracy value is 0.4800, F1-Score for the dirty gut disease category is 0.62, F1-Score for Other Diseases (Not Dirty Intestines) is 0.54, F1-Score for the Healthy Hands category is 0.29, and the overall F1-Score is 0.50. The white box test results show that the system can run well in all test scenarios applied. While the black box testing results show that the application functions as expected. In addition, the prediction results using the image import feature are supported by a confidence score with an average value of 48.89% for all three categories.
Implementation of EfficientNet-B0 CNN Model for Web-Based Strawberry Plant Disease Detection choirullah, Sultan; Yustanti, Wiyli
Journal of Emerging Information Systems and Business Intelligence (JEISBI) Vol. 6 No. 3 (2025): Vol. 06 Issue 03
Publisher : Universitas Negeri Surabaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26740/jeisbi.v6i3.72957

Abstract

Strawberry production in Indonesia has high economic value but is often hindered by plant diseases that reduce yield quality and quantity. Manual disease identification requires time, cost, and expertise, making it inefficient for farmers. This study proposes a web-based strawberry disease detection system by applying a Convolutional Neural Network (CNN) model using the EfficientNet-B0 architecture. The dataset consists of leaf, fruit, and flower images of strawberries in both healthy and infected conditions. The research followed the CRISP-DM framework, including business understanding, data preparation, modeling, evaluation, and deployment. The model was trained using transfer learning and fine-tuning techniques, with evaluation conducted through a confusion matrix and K-Fold Cross Validation. Experimental results indicate that the EfficientNet-B0 model achieved an overall accuracy of approximately 95.2% and demonstrated stable performance in classifying various strawberry plant diseases. The model achieved perfect accuracy (100%) in several classes such as Healthy Leaf, Leaf Spot, and Healthy Flower, while maintaining high accuracy in other classes like Fruit (95.2%) and Anthracnose Fruit Rot (94.7%), confirming its effectiveness in capturing essential visual features for accurate disease classification. The deployment of the model into a website using the Streamlit framework enables users to upload strawberry images and obtain automatic, fast, and accurate disease detection results. This system is expected to provide a practical solution to help farmers improve productivity and minimize losses caused by plant diseases.
CoAtNet for Chest X-Ray Report Generation with Bi-LSTM and Multi-Head Attention Akbar, Rafy Aulia; Putra, Ricky Eka; Yustanti, Wiyli
Indonesian Journal of Electronics, Electromedical Engineering, and Medical Informatics Vol. 7 No. 4 (2025): November
Publisher : Jurusan Teknik Elektromedik, Politeknik Kesehatan Kemenkes Surabaya, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35882/ijeeemi.v7i4.271

Abstract

In clinical environments, Chest X-Ray (CXR) represents the most prevalent diagnostic instrument, particularly facilitating diagnostic procedures through medical report. However, manual report preparation is time-consuming, highly dependent on the expertise of radiologists, and carries the risk of errors due to high workloads and limited expert staff. Therefore, an automated system based on artificial intelligence is needed to ease the workload of radiologists while increasing consistency. This study aims to develop an automated medical report generation system with balanced data distribution, reliable encoder, and bidirectional contextual understanding. The main contributions of this study include the implementation of an undersampling strategy based on majority captions followed by oversampling on minority labels while maintaining a proportion of labels with higher frequencies, the use of Bi-LSTM with Multi Head Attention (MHA) to strengthen text context understanding, and the use of CoAtNet as a visual encoder that combines the strengths of CNN and Transformer. The methodology incorporates image preprocessing via gamma correction for contrast improvement, data selection, balancing through combined undersampling and oversampling, and CoAtNet implementation as encoder paired with Bi-LSTM and MHA as decoder. Experimental execution employed the IU X-ray dataset, with assessment conducted using BLEU and ROUGE-L metrics. Outcomes revealed that the CoAtNet configuration with Bi-LSTM and MHA, coupled with the undersampling-oversampling strategy, delivered superior performance evidenced by a cumulative score of 1.642, with BLEU-1 to BLEU-4 and ROUGE-L achieving 0.480, 0.329, 0.245, 0.183, and 0.405, respectively. These findings prove that the combination of data balancing strategies with CoAtNet and Bi-LSTM is able to produce more accurate automated medical reports and reduce bias towards the majority label.
Optimizing Tuition Fee Determination with K-Means Cluster Relabeling Based on Centroid Mapping of Principal Component Pattern Yustanti, Wiyli; Iwan Nurhidayat, Andi; Iskandar Java, Muhammad
Journal of Information Systems Engineering and Business Intelligence Vol. 11 No. 3 (2025): October
Publisher : Universitas Airlangga

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.20473/jisebi.11.3.445-458

Abstract

Background: Tuition fee in Indonesian public universities is determined based on the socioeconomic status of prospective students. In this context, students are assigned to tuition fee groups after passing the selection process through achievement-based or computer-based exams. However, the current grouping system shows overlapping distributions, indicating the need for a more precise classification method.   Objective: This research aims to improve the accuracy of tuition fee group assignments by refining the clustering structure and relabeling the classification dataset.  Methods: A total of 13 socioeconomic variables were used to predict tuition fee groups. This research used K-Means clustering algorithm and a relabeling process using centroid mapping of principal components to balance original and newly generated labels. To assess the effectiveness of the relabeling process, six classification algorithms, namely Decision Tree (DT), K-Nearest Neighbors (KNN), Naive Bayes (NB), Logistic Regression (LR), Random Forest (RF), and Support Vector Machine (SVM), were used. Statistical tests at a 5% significance level were conducted to evaluate improvements in classification accuracy.  Results: The relabeling process significantly enhanced prediction accuracy compared to the original dataset. The refined clustering structure reported better classification performance across all six algorithms, showing the effectiveness of the proposed method.  Conclusion: The results showed that robust clustering and a relabeling method improved the precision of tuition fee classification systems. The proposed framework provided a data-driven solution for refining classification models, ensuring a fairer distribution of tuition fee based on socioeconomic indicators. The novelty lies in the centroid-based relabeling, which uses principal component patterns to enhance interpretability and classification accuracy. The method was adaptable for global use in any educational system using socioeconomic-based fee classification. Future research should explore alternative clustering methods and additional socioeconomic factors to enhance classification accuracy.    Keywords: K-Means Clustering, Machine Learning, Relabeling Process, Socioeconomic Indicators, Tuition Fee Classification   
Otomatisasi Klasifikasi Tingkat Urgensi Keluhan E-Layanan Unesa Berbasis TF-IDF dan Logistic Regression Alpiana, Intan; Yustanti, Wiyli; Yamasari, Yuni
Ranah Research : Journal of Multidisciplinary Research and Development Vol. 8 No. 1 (2025): Ranah Research : Journal Of Multidisciplinary Research and Development
Publisher : Dinasti Research

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.38035/rrj.v8i1.1912

Abstract

Perkembangan teknologi digital menuntut perguruan tinggi untuk menghadirkan layanan akademik yang cepat, tepat, dan responsif. Universitas Negeri Surabaya (Unesa) melalui platform E-Layanan memberikan sarana bagi civitas akademika untuk menyampaikan keluhan terkait kendala penggunaan sistem informasi dan jaringan. Namun, proses klasifikasi tingkat urgensi keluhan masih dilakukan secara manual oleh admin, yang berpotensi menyebabkan keterlambatan penanganan, inkonsistensi penilaian, serta meningkatnya beban kerja. Penelitian ini bertujuan untuk mengembangkan sistem otomatisasi klasifikasi tingkat urgensi keluhan dengan memanfaatkan Term Frequency-Inverse Document Frequency (TF-IDF) sebagai representasi fitur teks, serta Logistic Regression berbobot (class_weight) sebagai model klasifikasi utama. Dataset yang digunakan terdiri dari 79.303 keluhan, dibagi menjadi data latih (70%), validasi (15%), dan uji (15%). Evaluasi model dilakukan menggunakan metrik akurasi, presisi, recall, F1-score, dan confusion matrix. Hasil penelitian menunjukkan bahwa kombinasi TF-IDF dan Logistic Regression berbobot mampu memberikan kinerja yang baik dengan akurasi 92,54% pada data uji. Selain itu, model menunjukkan kemampuan yang tinggi dalam mendeteksi keluhan kritis secara akurat, memastikan prioritas penanganan terjaga secara optimal. Temuan ini menegaskan bahwa penerapan model berbasis pembelajaran mesin dapat meningkatkan efisiensi operasional dan konsistensi klasifikasi dibandingkan pendekatan manual. Sistem yang dikembangkan diharapkan dapat diintegrasikan lebih lanjut ke dalam platform E-Layanan Unesa, mendukung proses penanganan keluhan secara otomatis dan real-time, serta membantu administrasi fokus pada resolusi masalah yang paling mendesak.
PELATIHAN MEDIA PEMBELAJARAN MENGGUNAKAN CANVA UNTUK GURU MI AL AHMAD, KRIAN, SIDOARJO Naim Rochmawati; Yamasari, Yuni; Yustanti, WIyli; Qoiriah, Anita; Aviana, Anisah Nurul
Jurnal ABDI: Media Pengabdian Kepada Masyarakat Vol. 9 No. 1 (2023): JURNAL ABDI : Media Pengabdian Kepada masyarakat
Publisher : Universitas Negeri Surabaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26740/abdi.v9i1.19853

Abstract

Dalam proses belajar mengajar, media pembelajaran berperan penting. Media pembelajaran yang interaktif membantu para siswa lebih mudah dalam memahami konten materi yang disampaikan para guru. Dengan kesadaran untuk meningkatkan kemampuan dalam membuat media pembelajaran agar kualitas pembelajaran semakin meningkat, para guru MI Al Ahmad, Krian, Sidoarjo, meminta pelatihan pembuatan media pembelajaran. banyak tool yang bisa digunakan, salah satunya adalah Canva. Dalam Canva, disediakan banyak fasilitas menu untuk membuat media pembelajaran yang interaktif. Untuk itu, pelatihan kali ini adalah memberikan pelatihan Canva bagi para guru MI Al Ahmad untuk meningkatkan kemampuan digital para guru MI Al Ahmad dalam membuat media pembelajaran yang interaktif. Metode kegiatan adalah dengan model ceramah dilanjutkan dengan praktikum menggunakan Canva. Hasil dari pelatihan ini adalah kemampuan para guru MI Al Ahmad dalam membuat media pembelajaran interaktif menggunakan Canva. Dari hasil evaluasi kegiatan disimpulkan bahwa pelatihan ini dapat dikatakan berhasil meskipun masih perlu penyempurnaan dalam kegiatan yang dilakukan. Hal ini diindikasikan dengan respon yang diberikan oleh guru MI Al Ahmad, sebagai peserta pelatihan, pada angket online yang dibagikan setelah selesai pelatihan.
Pelatihan Pemanfaatan Internet untuk Menunjang Kreativitas Guru dalam Penyampaian Materi secara Daring Yamasari, Yuni; Qoiriah, Anita; Yustanti, Wiyli; Rochmawati, Naim; Nurhidayat, Andi Iwan; Kurniawan, Ari
Abdimas: Papua Journal of Community Service Vol. 6 No. 1 (2024): Januari
Publisher : Lembaga Pengembangan dan Pengabdian Masyarakat Universitas Muhammadiyah Sorong

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33506/pjcs.v6i1.2749

Abstract

   
A Hybrid Clustering–Classification Framework for SMEs Success Level Prediction Saputra, Andika Dermawan; Yustanti, Wiyli
JIEET (Journal of Information Engineering and Educational Technology) Vol. 9 No. 2 (2025)
Publisher : Universitas Negeri Surabaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26740/jieet.v9n2.p89-100

Abstract

Micro, Small, and Medium Enterprises (SMEs) are vital to economic growth, yet their complex success determinants necessitate advanced predictive modeling. This study proposes a hybrid clustering-classification framework to classify and predict SME success levels based on 22 multidimensional indicators, including financial literacy, FinTech adoption, and entrepreneurial resilience. K-Means clustering was first applied to the survey data, yielding three optimal success personas, validated by the highest Silhouette Score (0.5238). These clusters were labeled with Beginner and Conventional, Stable Digital Adopter, and Digital Innovator SMEs. These empirically derived clusters served as pseudo-labels for the classification stage. Classification algorithms were tested with and without the Synthetic Minority Oversampling Technique (SMOTE). While ensemble methods (Random Forest, LightGBM) and SVM performed well, the K-Nearest Neighbors (KNN) algorithm consistently outperformed all others, achieving the highest F1-Score (0.9324) under SMOTE implementation. The findings validate the effectiveness of the hybrid clustering-classification approach in accurately mapping and predicting SME success levels. The resulting model serves as a robust, data-driven tool for policymakers to guide targeted interventions and digital training programs, fostering sustainable SME development.
Optimizing UKT Prediction Based on Socio-Economic Features: A Multimodel Evaluation with Feature Selection Srategies Putri, Windy Chikita Cornia; Yustanti, Wiyli; Yohannes, Ervin
EDUTIC Vol 12, No 2: 2025
Publisher : Universitas Trunojoyo Madura

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21107/edutic.v12i2.31828

Abstract

Determining the tuition fee group (UKT) for new students in Indonesian public universities represents a complex challenge requiring an equitable, data-driven approach. This study introduces an integrative feature selection strategy that combines five popular techniques Chi-Square, Recursive Feature Elimination (RFE), LASSO Regression, Random Forest Importance, and Exploratory Factor Analysis (EFA) to extract the most relevant attributes from 53 socioeconomic variables of prospective students at Universitas Negeri Surabaya. As a novelty, the study identifies intersecting features consistently selected by all five methods and evaluates their impact on the performance of five classification algorithms: Support Vector Machine (SVM), Decision Tree, Random Forest, K-Nearest Neighbor (KNN), and Naïve Bayes. Experimental results demonstrate a significant improvement in accuracy, with SVM increasing from 0.7550 to 0.7810. These findings confirm that integrative feature selection can optimize model performance while reducing data complexity. This study provides a replicable methodological contribution for developing transparent and adaptive classification systems based on socioeconomic data in higher education contexts.