Claim Missing Document
Check
Articles

Found 17 Documents
Search

Klasifikasikan Jenis Cacat Kulit Menggunakan SMOTE-GoogLeNet Prananda, Alifia Revan; Frannita, Eka Legya; Pramitaningrum, Erlita; Hidayat, Anwar; Setiawan , Wawan Budi; Purwaningsih , Nunik
JITU : Journal Informatic Technology And Communication Vol. 8 No. 1 (2024)
Publisher : Universitas Boyolali

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36596/jitu.v8i1.1341

Abstract

Deep learning has been proven to be able to provide significant contributions to several fields, including industry. It has also been proven that it has resulted in an outstanding performance for classification, detection, and even segmentation processes. In the leather industry, it also successfully gave valuable results, especially for the leather defect inspection process. However, despite its outstanding performance, it remained a drawback because it produced insignificant results if employed in a small or imbalanced dataset. This research work focuses on the analysis of the implementation of the data balancing method for improving the performance of the deep learning method for classifying the types of leather defects. This research work was done by employing three processes. In the first step, we utilized the data balancing method to balance the data proportion. In the next step, we employed GoogLeNet as a deep learning architecture for training and testing processes. Our experiment was conducted in two scenarios. The first scenario was done by using the original dataset. Whereas the second scenario was accomplished by utilizing the data balancing method before training and testing. According to the experiment results, implementing the data balancing method successfully increased the performance of the deep learning method by more than 15%. It can be inferred that the proportion or the number of data strongly affected the performance of deep learning models.
Pemberdayaan Keamanan Dan Kesejahteraan Melalui Penggunaan CCTV, Digital Marketing, dan Website Ibrahim Nawawi; Rheza Ari Wibowo; D Jayus Nor Salim; Suamanda Ika Novichasari; Imam Adi Nata; Damar Wicaksono; Alifia Revan Prananda; Restu Rakhmawati
Ngudi Waluyo Empowerment: Jurnal Pengabdian Kepada Masyarakat Vol. 2 No. 2 (2023): Ngudi Waluyo Empowerment: Jurnal Pengabdian Kepada Masyarakat
Publisher : Fakultas Komputer dan Pendidikan Universitas Ngudi Waluyo

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

Abstract

Desa Sidogede, yang terletak di Kecamatan Grabag, Kabupaten Magelang, memiliki potensi besar untuk meningkatkan keamanan dan kesejahteraan warganya. Proyek "Bakti Sosial Pemberdayaan Keamanan dan Kesejahteraan" bertujuan untuk mengintegrasikan teknologi CCTV, Digital Marketing, dan Website Desa Interaktif sebagai solusi holistik untuk meningkatkan kondisi sosial dan ekonomi di desa ini. Penggunaan CCTV akan memperkuat keamanan dengan memonitor aktivitas publik dan rumah-rumah warga. Digital Marketing akan membantu mendukung pengembangan usaha lokal dan mempromosikan produk-produk desa. Sementara itu, Website Desa Interaktif akan menjadi sumber informasi penting dan sarana partisipasi warga dalam pengambilan keputusan. Proyek ini tidak hanya bertujuan untuk meningkatkan tingkat keamanan fisik, tetapi juga ekonomi desa melalui promosi produk lokal dan peningkatan akses ke informasi. Dengan demikian, proyek ini diharapkan akan berdampak positif terhadap kesejahteraan dan kehidupan masyarakat Desa Sidogede, menjadikannya sebagai contoh terbaik bagi upaya pemberdayaan desa-desa di wilayah sekitarnya.
Hidroponik dengan Konsep Agro Voltec Guna Optimalisasi Kemandirian dan Ketahanan Pangan di Kelurahan Rejowinangun Selatan Kota Magelang Alifia Revan Prananda; Cornelius Rangga Surya Kusuma; Daniel Gunawan; Terra Rhebekka; Galih Slamet
Manfaat : Jurnal Pengabdian Pada Masyarakat Indonesia Vol. 1 No. 4 (2024): November : Manfaat : Jurnal Pengabdian Pada Masyarakat Indonesia
Publisher : Asosiasi Riset Ilmu Tanaman Dan Hewan Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62951/manfaat.v1i4.113

Abstract

Increasing population density and concretization in urban areas, such as Magelang City, have caused a significant decrease in green open spaces (RTH), which has an impact on food security and independence. Magelang City, with a population density of 7,361 people/km², faces a major challenge in providing adequate green spaces to support local food production. Rejowinangun Selatan Village, as one example, shows the limited green open spaces available. To overcome this problem, urban farming, especially through hydroponic techniques, offers an effective solution. The methods applied in this community service include various approaches to educate and directly involve the community in the practice of making hydroponics with solar panels. By utilizing renewable energy-based technology, such as solar panels and integrated electrical systems, hydroponic farming allows food production in limited spaces such as house terraces. This system not only increases food independence and security, but also ensures that the vegetables produced are healthier because of the minimal use of pesticides. Urban farming with the hydroponic method can optimize the use of limited space and become a productive and sustainable alternative for urban communities.
Using Support Vector Machines for Predicting and Mitigating Stunting in Early Childhood Education in Rural Semarang Novichasari, Suamanda Ika; Prananda, Alifia Revan; Suwidagdho, Dhanang; Fauziah, Syifa; Setya Wijaya, Vania Amelia; Adam, Otmar Shah
Jurnal Obsesi : Jurnal Pendidikan Anak Usia Dini Vol. 8 No. 5 (2024)
Publisher : LPPM Universitas Pahlawan Tuanku Tambusai

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31004/obsesi.v8i5.6131

Abstract

In 2030, 70% of Indonesia's population will be of productive age (15-64 years), which is a demographic bonus. However, this potential is threatened by the high rate of stunting in children, which threatens future workforce productivity. Early identification of stunting risk is essential for timely intervention. This study develops a stunting prediction model using machine learning with data from early childhood education institutions in rural Semarang. The model used is a Support Vector Machine (SVM) implemented through the RapidMiner framework. The SVM model achieved an accuracy of 97.56%, a precision of 98.97%, a recall of 97.37%, and an AUC of 0.997. The results of the SVM model highlight the importance of physical motor skills and artistic development.
Development and Validation of the Vark Instrument in Citizenship Education Learning: A Structural Equation Modeling Approach for More Accurate Learning Style Measurement Widiyanto, Delfiyan; Novitasari, Novitasari; Syahroni, Mashud; Prananda, Alifia Revan; Mujtahidah, Tholibah; Istiqomah, Annisa
Journal Evaluation in Education (JEE) Vol 6 No 2 (2025): April
Publisher : Cahaya Ilmu Cendekia Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.37251/jee.v6i2.1561

Abstract

Purpose of the study: This research aims to develop and test the VARK learning style instrument in Citizenship Education Learning using SEM (Structural Equation Modeling) to produce a valid, reliable and goodness of fit instrument for more accurate measurement of learning styles. Methodology: This study employs quantitative methods to evaluate the VARK learning styles theory in citizenship education through three primary phases: (1) assessing interrater validity to verify the construct's feasibility, (2) establishing empirical validity via Confirmatory Factor Analysis (CFA) to examine the construct's effectiveness and reliability, and (3) utilizing Structural Equation Modeling (SEM) along with goodness-of-fit (GOF) analysis to determine the model's overall suitability. Main Findings: The findings of this research indicate that the VARK learning styles framework and related instruments in citizenship education are very suitable for measurement. The model meets the main criteria: (1) Inter-rater Validity Test confirms its validity, (2) Construct Validity Test using Confirmatory Factor Analysis shows validity and reliability, and (3) Goodness-of-Fit Test places the model in the acceptable fit category. So this instrument can be used to measure learning styles accurately. Novelty/Originality of this study: The novelty of this research lies in the development and instrument of VARK learning styles in Citizenship Education learning which was tested using Structural Equation Modeling (SEM) to produce a feasible and accurate instrument for measuring learning styles.
Penerapan Faster RCNN + ResNet 50 untuk Mengidentifikasi Spesies dan Stadium Parasit Plasmodium Malaria Prananda, Alifia Revan; Novichasari, Suamanda Ika; Fatkhurrozi, Bagus; Abdillah, Muhammad Nurkholis; Frannita, Eka Legya; Majidah, Zharifa Nur; Wibowo, Fadhila Syahida
TIN: Terapan Informatika Nusantara Vol 6 No 2 (2025): July 2025
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/tin.v6i2.7187

Abstract

Malaria is one of the epidemic health diseases and is well-known as a serious infectious disease. The malaria examination process had occurred by analyzing the digital microscopic images using a microscope. Those examination procedures were conducted manually, which lead to some hurdles such as misinterpretation, misdiagnosis and may produce subjective results. This research aims to develop a method for detecting the Plasmodium parasite and identifying the species and stage of Plasmodium parasite. The proposed method was performed into 488 raw data comprising of 538 parasites. The proposed method was started by conducting a data augmentation process for balancing the number of data, training model, testing model, evaluation. In this study, both the training and testing processes were performed by applying Faster RCNN + ResNet-50. The result of the testing process shows that Faster RCNN + ResNet-50 successfully achieved mAP of 0,603. It also achieved accuracy of 93.91%, sensitivity of 66.20%, specificity of 96.10%, PPV of 60.14% and NPV of 97.30%. This result indicates that the proposed method is powerful for detecting Plasmodium parasites and identifying all species and stadiums.
Pengembangan Intelligent Leather Inspection Method Berbasis Interpretable Artificial Intelligence Frannita, Eka Legya; Wulandari, Dwi; Putri, Naimah; Rahmawati, Atiqa; Prananda, Alifia Revan
TIN: Terapan Informatika Nusantara Vol 6 No 2 (2025): July 2025
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/tin.v6i2.7425

Abstract

The Industry 4.0 revolution, characterized by the widespread adoption of artificial intelligence and automation, has fundamentally transformed quality inspection processes in manufacturing sectors. Nevertheless, the leather tanning industry continues to rely on conventional visual inspection methods conducted by human operators, which are inherently susceptible to subjectivity, inter-operator variability, and inconsistent outcomes. This study proposes an integrated deep learning framework utilizing the NasNet-Large architecture combined with Local Interpretable Model-Agnostic Explanations (LIME) to automate objective defect detection and quality classification of pickled leather. The research employs a digital image dataset comprising four distinct leather grade categories, each annotated with expert-validated ground truth labels and professional interpretations. Experimental results demonstrate consistent model performance with 75% accuracy in both training and validation phases while achieving improved testing accuracy of 79%. LIME-based interpretability analysis reveals significant spatial convergence between model-identified defect regions and expert-annotated ground truth references. These findings indicate that the developed model exhibits remarkable competence in replicating professional leather quality inspection capabilities. The proposed approach not only enhances inspection efficiency by reducing human-dependent errors but also provides transparent decision-making interpretability - a critical requirement for reliable AI implementation in industrial applications. This research contributes to the advancement of explainable AI systems in material quality assessment, offering methodological innovation and practical implementation value for the leather manufacturing sector.