Article Per Year (5 Year)
p-Index From 2021 - 2026
3.044
P-Index
This Author published in this journals
All Journal ELKHA : Jurnal Teknik Elektro Indonesian Journal on Software Engineering (IJSE) JOURNAL OF APPLIED INFORMATICS AND COMPUTING Indonesian Journal of Electrical Engineering and Computer Science Abdimasku : Jurnal Pengabdian Masyarakat Journal of Intelligent Computing and Health Informatics (JICHI) Community: Jurnal Pengabdian Pada Masyarakat Jurnal Informatika: Jurnal Pengembangan IT Engagement: Jurnal Pengabdian Kepada Masyarakat Scientific Journal of Informatics
Paramita, Cinantya
Unknown Affiliation
Religion Arts Humanities Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Dentistry Economics, Econometrics & Finance Education Electrical & Electronics Engineering Energy Engineering Health Professions Industrial & Manufacturing Engineering Languange, Linguistic, Communication & Media Medicine & Pharmacology Public Health Social Sciences

Published : 22 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 22 Documents
Search

 1   2   3 
Deep Learning-Based Eye Disorder Classification: A K-Fold Evaluation of EfficientNetB and VGG16 Models Paramita, Cinantya; Rakasiwi, Sindhu; Andono, Pulung Nurtantio; Shidik, Guruh Fajar; Shier Nee Saw; Rafsanjani, Muhammad Ivan
Scientific Journal of Informatics Vol. 12 No. 3: August 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i3.26257

Abstract

Purpose: The study evaluates EfficientNetB3 and VGG16 deep learning architectures for image classification, focusing on stability, accuracy, and interpretability. It uses Gradient-weighted Class Activation Mapping to improve transparency and robustness. The research aims to create reliable AI-based diagnostic tools. Methods: The study used a dataset of 4,217 color retinal fundus images divided into four classes: cataract, diabetic retinopathy, glaucoma, and normal. The dataset was divided into 70% for training, 10% for validation, and 20% for testing. The researchers used a transfer learning approach with EfficientNetB3 and VGG16 models, pretrained on ImageNet. Real-time augmentation was applied to prevent overfitting and improve generalization. The models were compiled with the Adam optimizer and trained with categorical cross-entropy loss. Early stopping was implemented to allocate computational resources efficiently and reduce overfitting. A learning rate scheduler (ReduceLROnPlateau) was added to adjust the learning rate if no significant improvement was made concerning validation loss. EfficientNetB3 was more efficient in model size, possessing only 12 million parameters compared to VGG16's 138 million, making it suitable for resource-constrained mobile or embedded systems. The final evaluation was done on the held-out test set. Result: The EfficientNetB3 architecture outperforms VGG16 in classification accuracy and loss value stability, with an average accuracy of 93%. It also exhibits better transparency and predicted accuracy, making it a reliable model for medical image categorization. Novelty: This work introduces a novel framework integrating EfficientNetB3 architecture, stratified cross-valuation, L2 regularization, and Grad-CAM-based interpretability, focusing on openness and explainability in model evaluation.
Analisis Pengaruh SMOTE terhadap Kinerja Model KNN untuk Prediksi Risiko Stroke Paramita, Cinantya; Simbolon, Calvin Samuel; Pamungkas, Azriel Sebastian; Triono, Justin Matthew; Widi Utomo, Emanuel Pinesthi; Subhiyakto, Egia Rosi
Jurnal Informatika: Jurnal Pengembangan IT Vol 10, No 4 (2025)
Publisher : Politeknik Harapan Bersama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30591/jpit.v10i4.8809

Abstract

Penelitian ini membahas masalah ketidakseimbangan data dalam klasifikasi risiko stroke, di mana kasus non-stroke secara signifikan lebih rendah daripada kasus stroke. Ketidakseimbangan kelas cenderung menimbulkan bias terhadap kelas mayoritas, yang menyebabkan berkurangnya efektivitas klasifikasi. Untuk mengatasi hal ini, SMOTE (Synthetic Minority Over-sampling Technique) digunakan untuk mengatasi ketidakseimbangan kelas dalam dataset dan algoritma K-Nearest Neighbor (KNN) digunakan untuk klasifikasi. Dataset mengalami preprocessing, aplikasi SMOTE, dan algoritma KNN dilatih dan dievaluasi menggunakan metrik standar termasuk akurasi, presisi, recall, dan F1-score. Penerapan SMOTE bersama dengan KNN menghasilkan peningkatan yang signifikan dalam hasil klasifikasi, mencapai akurasi 91,87%, presisi 94,27%, recall 89,20%, dan F1-score 91,66%. Temuan ini menegaskan bahwa pendekatan yang diimplementasikan berkinerja baik dalam mendeteksi risiko stroke meskipun ada set data yang tidak seimbang. Tujuan dari penelitian ini adalah untuk menginformasikan kemajuan teknologi deteksi dini stroke yang lebih kuat dan mendukung peningkatan dalam penyediaan layanan kesehatan.
Pelatihan Penggunaan Aplikasi Screen Reader JAWS Bagi Tunanetra Untuk Meningkatkan Kemampuan Dalam Pengelolaan Administrasi Paramita, Cinantya; Sudibyo, Usman; Muljono, Muljono; Supriyanto, Catur
ABDIMASKU : JURNAL PENGABDIAN MASYARAKAT Vol 2, No 2 (2019): Juli 2019
Publisher : LPPM UNIVERSITAS DIAN NUSWANTORO

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (586.68 KB) | DOI: 10.33633/ja.v2i2.46

Abstract

Salah satu permasalahan yang dihadapi Perkumpulan Penyandang Disabilitas Indonesia (PPDI), Dewan Pengurus Cabang (DPC) Kota Semarang yakni terbatasnya media dan prasarana untuk mendukung kegiatan belajar mereka dalam mengoperasikan komputer serta tingkat perekonomian yang hanya cukup untuk memenuhi kebutuhan hidup, meraka pun belum sepenuhnya paham dalam perkembangan teknologi hingga sampai saat ini semua masih diolah dalam bentuk manual yakni melaporkan dengan lisan dan dengan cara mengingat. Untuk mengatasi permasalahan tersebut, pengabdian ini mengusulkan untuk mengadakan pelatihan penggunaan Job Access with Speech (JAWS) bagi para tuna netra. Pengabdian dilakukan dengan mengajarkan penggunaan dasar keyboard yang didukung oleh aplikasi JAWS dan pengetahuan dasar Microsoft Excel.
Mitigating Class Imbalance in DDoS Detection: The Impact of Random Over Sampling on Machine Learning Performance Ghozi, Wildanil; Hussein, Jasim Nadheer; Sani, Ramadhan Rakhmat; Rafrastara, Fauzi Adi; Paramita, Cinantya; Supriyanto, Catur
ELKHA : Jurnal Teknik Elektro Vol. 17 No.2 October 2025
Publisher : Faculty of Engineering, Universitas Tanjungpura

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.26418/elkha.v17i2.95037

Abstract

Distributed Denial of Service (DDoS) attacks are a major cybersecurity threat, involving malicious traffic generated from numerous compromised sources to overwhelm and disable targeted services. Although machine learning (ML) has shown promise in detecting DDoS attacks through network traffic analysis, a key challenge remains: the class imbalance in datasets such as UNSW-NB15, where normal traffic significantly outweighs attack instances. This imbalance leads to biased predictions and degraded detection performance for minority attack classes. To address this issue, our study investigates the impact of Random Over Sampling (ROS), a simple yet effective balancing technique on improving detection accuracy in multi-class DDoS classification tasks. While prior works have primarily focused on ensemble algorithms or feature selection, our approach is distinct in emphasizing the effect of data balancing on macro evaluation metrics such as macro precision, macro recall, and macro F1-score. ROS was selected over more complex alternatives, such as SMOTE or ADASYN, due to its computational efficiency and ability to establish a performance baseline without introducing synthetic noise. We evaluate four machine learning algorithms: Decision Tree, Naïve Bayes, Random Forest, and XGBoost, using the UNSW-NB15 dataset. The results show that Decision Tree combined with ROS yields the highest improvement in macro F1-score, increasing by 36%. However, this improvement is accompanied by a moderate reduction in accuracy for certain algorithms. These findings highlight the critical role of class balancing in enhancing the reliability of DDoS detection models, especially in imbalanced multi-class scenarios.
Menavigasi Dunia Digital dengan Meningkatkan Literasi Office, TI, dan Internet di Kalangan Siswa-Siswi Pondok Pesantren Raudhatul Qur'an Paramita, Cinantya; Andono, Pulung Nurtantio; Sudibyo, Usman; Rafrastara, Fauzi Adi; Supriyanto, Catur
ABDIMASKU : JURNAL PENGABDIAN MASYARAKAT Vol 6, No 2 (2023): Mei 2023
Publisher : LPPM UNIVERSITAS DIAN NUSWANTORO

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33633/ja.v6i2.1338

Abstract

Peningkatan popularitas penggunaan perangkat komputer semakin berkembang di berbagai lapisan masyarakat. Pondok pesantren, yang sebelumnya dianggap sebagai tempat yang kurang produktif dan hanya diperuntukkan bagi mereka yang beragama, kini melakukan inovasi untuk meningkatkan peran dan potensi dalam mendukung kemaslahatan lingkungan sekitarnya. Pondok Pesantren Raudhatul Qur’an di Kauman Semarang telah banyak menciptakan siswa yang berhasil menghafal Al-Quran. Setelah menyelesaikan studi di pondok, banyak dari mereka yang melanjutkan pendidikan ke sekolah formal atau menjadi pemuka agama yang memberikan pengajaran dan bimbingan kepada masyarakat dalam memahami agama Islam di lingkungan mereka. Oleh karena itu, pelatihan teknologi komputer diperlukan untuk memberikan pengetahuan dan keterampilan bagi para santri agar dapat dimanfaatkan untuk membantu mengurus keperluan administrasi di pondok pesantren dan berguna bagi masa depan mereka. Sebanyak 53 santri diikutsertakan untuk mengikuti pelatihan yang mencakup pengenalan dasar teknologi informasi [1] seperti hardware, software, penggunaan aplikasi office seperti Word, Excel, dan PowerPoint, serta internet untuk komunikasi dan pengiriman data digital. Berdasarkan hasil pelatihan yang dilaksanakan, para santri memberikan respon positif seperti yang terlihat pada diagram 3 dan 4. Pada diagram 3 menunjukkan bahwa 81,4% dari para santri sangat tertarik dengan pelatihan tersebut, sementara hanya 13,9% yang merasa biasa-biasa saja dan 10,7% yang terpaksa mengikuti. Selain itu, hasil perbandingan pretest dan postest pada diagram 4 menunjukkan peningkatan yang signifikan setelah para santri mengikuti pelatihan tersebut.
Pemanfaatan Algoritma K-Means untuk Membuktikan Implementasi Undang-Undang Pelanggaran Hukum Korupsi di Pengadilan Negeri Banjarmasin Paramita, Cinantya; Rafrastara, Fauzi Adi; Supriyanto, catur
Jurnal Informatika: Jurnal Pengembangan IT Vol 8, No 2 (2023)
Publisher : Politeknik Harapan Bersama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30591/jpit.v8i2.5216

Abstract

This research aims to demonstrate the implementation of the Anti-Corruption Law in the Banjarmasin District Court by utilizing the K-Means algorithm. Corruption, which persists in Indonesia over a prolonged period, has reached a critical level, making it crucial to enforce the law fairly and firmly. In this study, the panel of judges in the Banjarmasin District Court was analyzed using the K-Means Clustering method and silhouette coefficient to decide corruption cases that result in state losses. The research findings indicate that the optimal number of clusters is 3, with a value of 0.686. However, there is also a lowest value among the 4 clusters, which is 0.454. These clusters are then divided into three categories of enforcement, namely cases that have been executed (108 cases), cases that will be executed (26 cases), and cases that have not been executed (2 cases). All clusters have a silhouette score of 0.742, indicating successful enforcement. This research provides concrete evidence that the panel of judges in the Banjarmasin District Court has implemented the Anti-Corruption Law while considering state losses. By utilizing the K-Means algorithm, this study also contributes to a better understanding of enforcement practices in the court. It is expected that the results of this research will support efforts to enhance the implementation of the Anti-Corruption Law in Indonesia, particularly in the Banjarmasin District Court
Deteksi Malware menggunakan Metode Stacking berbasis Ensemble Rafrastara, Fauzi Adi; Supriyanto, Catur; Paramita, Cinantya; Astuti, Yani Parti
Jurnal Informatika: Jurnal Pengembangan IT Vol 8, No 1 (2023)
Publisher : Politeknik Harapan Bersama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30591/jpit.v8i1.4606

Abstract

Serangan malware kian hari kian memprihatinkan. Evolusi malware yang cepat dan semakin destruktif menimbulkan kekhawatiran bagi banyak pihak. Oleh karena itu, deteksi malware yang efektif sangat dibutuhkan. Data mining memainkan peran yang krusial dalam bidang ini, mengingat algoritma-algoritma yang ada pada data mining bisa dilatih hingga menghasilkan akurasi yang paling tinggi. Untuk mengklasifikasi suatu file, apakah tergolong malware atau tidak, dalam penelitian ini metode stacking digunakan karena dapat meningkatkan akurasi jika dibandingkan dengan algoritma-algoritma klasifikasi konvensional. Empat Algoritma dilibatkan dalam eksperimen yang dilakukan, yaitu: Neural Network, Random Forest, kNN, dan Logistic Regression. Tiga algoritma pertama digunakan sebagai classifier pada level 0, sementara itu Logistic Regression digunakan classifier pada level 1 (meta classifier).  Dengan kombinasi 4 algoritma tersebut, akurasi yang diperoleh adalah sebesar 98.7%, dan akurasi tersebut merupakan yang paling tinggi jika dibandingkan dengan masing-masing algoritma jika dieksekusi secara individual.
Aplikasi Prediksi IHSG Berbasis Web Dengan Integrasi Multi-Algoritma Waluyo, Dwi Eko; Paramita, Cinantya; Kinasih, Hayu Wikan; Pergiwati, Dewi; Rafrastara, Fauzi Adi
Jurnal Informatika: Jurnal Pengembangan IT Vol 9, No 2 (2024)
Publisher : Politeknik Harapan Bersama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30591/jpit.v9i2.6193

Abstract

The four regression algorithms used in predicting the Composite Stock Price Index (IHSG) have contributed significantly, as the test results show that the Decision Tree algorithm outperforms k-Nearest Neighbor, Linear Regression, and Random Forest, especially in terms of Mean Squared Error (MSE) and R2 score. The stages of data collection, pre-processing, and modeling, followed by model performance measurement, have provided valuable insights into the effectiveness of each algorithm. The success of the Decision Tree in this testing has further propelled its development into a web-based application. This conversion process, following the outlined flowchart, integrates various essential aspects of the model, including user interface and back-end integration, ensuring that the application can be accessed and used efficiently and effectively. Furthermore, the black box testing and User Acceptance Testing (UAT) results, using the Mean Opinion Score (MOS), enhance the validity and reliability of the application. Black box testing involving 2 features with 37 steps demonstrates the system's effectiveness in producing valid outputs, from the initial menu display to the prediction results. Additionally, UAT involving students and entrepreneurs as respondents provides in-depth insights into user acceptance. With a focus on functionality at 97.08%, reliability at 96.09%, and usability at 98.09%, UAT yields high scores in all three aspects, with usability achieving the highest score. These results not only confirm the efficiency of the system in performing its functions but also indicate a high level of user satisfaction, strongly suggesting the potential for widespread adoption of this application in the future.
Pengembangan Sistem Klasifikasi Karakteristik Siswa Berbasis Website dengan menggunakan Algoritma C4.5 Paramita, Cinantya; Rafrastara, Fauzi Adi; Kencana, Lisdi Inu
Jurnal Informatika: Jurnal Pengembangan IT Vol 8, No 1 (2023)
Publisher : Politeknik Harapan Bersama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30591/jpit.v8i1.4678

Abstract

Student characteristics are one of the attributes of knowing a student's thinking skills and academic abilities. In the process of teaching and learning, appropriate learning strategies must be applied to students. The Hippocrates-Galenus typology categorizes personality types into four different categories, namely sanguine, choleric, melancholic and phlegmatic. Classification of characteristics that use an approach to students based only on experience or intuition can produce inaccurate results and take a lot of time to process. A system with the ability to predict student characteristics is needed in order to be able to assess students more quickly. In this study, the C4.5 algorithm was implemented into a system that aims to carry out the process of classifying the characteristics of students. From the results of the tests carried out, the C4.5 algorithm obtains an accuracy of 90.08%. This shows it is able to classify student characteristics well by using the C4.5 algorithm
Komparasi dan Implementasi Algoritma Regresi Machine Learning untuk Prediksi Indeks Harga Saham Gabungan Waluyo, Dwi Eko; Kinasih, Hayu Wikan; Paramita, Cinantya; Pergiwati, Dewi; Nohan, Rajendra; Rafrastara, Fauzi Adi
Jurnal Informatika: Jurnal Pengembangan IT Vol 9, No 1 (2024)
Publisher : Politeknik Harapan Bersama

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30591/jpit.v9i1.6105

Abstract

Indeks Harga Saham Gabungan (IHSG) or Indonesia Composite Index (ICI) is part of the macro indicators of a country that describes the economic condition of a country. ICI is an interesting study to research since its existence will be able to show market sentiment regarding an event that occurred in a country. This research tries to predict the ICI in the future based on historical data. The dataset used in this research is publicly available in Yahoo Finance. The experiment is conducted by implementing some regression machine learning algorithms, such as Decision Tree, Random Forest, k-Nearest Neighbor (kNN), and Linear Regression. As a result, Decision Tree has the lowest MSE value compared to other methods: 1268.242. In this research, a website-based application prototype was also developed that can be used to view IHSG graphs and make future predictions, using the 4 (four) tested algorithms.
Co-Authors Abu Salam Adhitya Nugraha Adi Nugroho Ahmed, Foez Amalia Amalia Andreas Wilson Setiawan Astuti, Yani Parti Catur Supriyanto Dwi Eko Waluyo Dwi Nurul Izzhati Dzaky, Azmi Abiyyu Egia Rosi Subhiyakto, Egia Rosi Erwin Yudi Hidayat Fauzi Adi Rafrastara Febrian, Nanda Dwi Finoe, Albireo Musyaffa Guruh Fajar Shidik Hayu Wikan Kinasih Hussein, Jasim Nadheer Junta Zeniarja Kencana, Lisdi Inu Manaf, Humam Wajdi Muljono Muljono Mulyanto, Edy Nohan, Rajendra Octaviani, Dhita Aulia Pamungkas, Azriel Sebastian Pergiwati, Dewi Pradana, Naufal Arsyaputra Pulung Nurtantio Andono Putra Hartanto, Octaviano Ryan Eka Rafsanjani, Muhammad Ivan Rahadian, Arief Ramadhan Rakhmat Sani Rodia Syamwil Saputra, Filmada Ocky Savicevic, Anamarija Jurcev Shier Nee Saw Simbolon, Calvin Samuel Sindhu Rakasiwi Sudibyo, Usman Swanny Trikajanti Widyaatmadja Syamsiar, Syamsiar Tanjung, Mohamad Aprilian Tasleem, Faathir El Tjahyono, Rudi Triono, Justin Matthew Widi Utomo, Emanuel Pinesthi Wildanil Ghozi ‘Aini, Fariida Qurrota