LPPI UAD

30/10/2024

https://youtu.be/YuR0r-E7FJI
The study proposes an improved Random Forest (RF) algorithm based on hierarchical clustering (HCRF). HCRF optimizes feature selection by establishing similar feature groups based on the GINI index and selecting features proportionally to construct a feature subset. This process reduces the impact of useless and redundant features, improving the model's generalization ability and overall performance. The HCRF algorithm showed significant improvements in all evaluation indicators, making it superior to other classifiers.



Master Program of Electrical Engineering (MEE), UAD
https://mee.uad.ac.id/

Authors: Wang Zhuo, Azlin Ahmad (IJEECS ID 38396)Random Forest (RF) selects feature subsets randomly. Useless and redundant features will lower the quality ...

27/10/2024

https://youtu.be/zl6hJPscZGg
The modern power system is increasingly relying on solar photovoltaic power plants (PV PPs) to replace conventional synchronous generators. However, these plants lack natural rotational inertia, leading to frequency instability. To address this, the researchers limit PV PP pe*******on and provide virtual inertia from energy storage systems. They also propose a strategy to optimize the Fast Frequency Response (FFR) capacity of deloaded PV PPs using the Particle Swarm Optimization algorithm.



Magister of Electrical Engineering (MEE), UAD
https://mee.uad.ac.id

Optimal Deployment of Solar PV power plants as fast frequency response source for a frequency secure low inertia power gridAuthors: Brian K. Wamukoya, Christ...

26/10/2024

https://youtu.be/h_Z7yH13Fe8
This article explores reducing grounding resistance in dry lands, specifically in acidic swamps. It uses a cost-effective dolomite mixture to address high soil resistivity. The study found that dolomite can generally overcome grounding resistance in acidic swamps, but the decrease was inconsistent due to changes in dolomite composition and rod electrode material. The method offers a range of resistance prices.



S2 Teknik Elektro UAD https://mee.uad.ac.id/

Authors: Dian Eka Putra, Muhammad Irfan Jambak, Zainuddin Nawawi (TELK 26656)The study of reducing grounding resistance on dry or sandy land generally employ...

26/10/2024

https://youtu.be/AQG3YWMh7cs
This study evaluates the effectiveness of various term frequency-inverse document frequency (TF-IDF) modifications for short text classification, addressing imbalanced datasets. Combining standard, log-scaled, and boolean TF-IDF with oversampling methods yields the best results, with oversampling methods outperforming undersampling methods in all experiments.


S2 Teknik Elektro UAD https://mee.uad.ac.id/

Comparing Oversampling and Undersampling in Different TF-IDF Vectorizers for Extremely Imbalanced Indonesian Language Short Text ClassificationAuthors: I Nyo...

26/10/2024

S2 Teknik Elektro UAD
https://youtu.be/teHuza7HkUA
This paper presents a flexible design of patch antenna arrays (PAA) to address bandwidth limitations in the 2.4 GHz Industrial, Scientific, and Medical (ISM) band. The authors designed and evaluated 1x2, 2x2, and 2x4-element rectangular PAAs using a quarter-wave transformer method and a parallel scheme for connecting patches. They used FR-4 epoxy as the substrate material and conducted extensive simulations using Ansys HFSS. The 2x4 PAA had excellent performance metrics, including a gain of 9.17 dB, directivity of 13 dB, and a voltage standing wave ratio below 0.5.

Authors: Kymbat Kopbay, Madiyar Nurgaliyev, Ahmet Saymbetov, Nurzhigit Kuttybay, Askhat Bolatbek, Sayat Orynbassar, Batyrbek Zholamanov (TELK 26344)Research ...

24/10/2024

Electro-capacitive cancer therapy using wearable electric field detector: a review
DOI: https://doi.org/10.11591/csit.v5i3.p292-305

ECCT, or Electro-capacitive Cancer Therapy, is a non-invasive, personalized approach to cancer treatment using wearable electric field detectors. These devices detect changes in electric fields, providing real-time monitoring for personalized treatment plans. The safety and efficacy of electric field exposure in vital organs are being investigated, contributing to improved diagnostic techniques and safety measures in medical and engineering fields. This technology enhances personalized medicine, improving treatment outcomes and patient quality of life.

Electro-capacitive cancer therapy using wearable electric field detector: a review

24/10/2024

Capabilities of cellebrite universal forensics extraction device in mobile device forensics
DOI: https://doi.org/10.11591/csit.v5i3.p254-264

Cellebrite UFED is a digital forensics tool that aids investigators in solving criminal and cybersecurity cases by extracting and analyzing mobile device data. It can recover data from erased or obscured devices, including call logs, texts, emails, and social media. The intuitive user interface speeds up data extraction, revealing crucial information. Legal and ethical considerations are crucial in mobile device forensics, with AI and ML algorithms potentially automating data extraction in future tools.

Capabilities of cellebrite universal forensics extraction device in mobile device forensics

24/10/2024

Securing DNS over HTTPS traffic: a real-time analysis tool
DOI: https://doi.org/10.11591/csit.v5i3.p227-234

A study developed a live tool to analyze DNS over HTTPS (DoH) traffic and classify it as benign or malicious. Using machine learning algorithms like K-NN, RF, DT, DNN, and SVM, the tool achieved exceptional performance. It integrates with the Mallory tool for local DNS resolution, allowing for more accurate simulation of DoH queries. The tool's effectiveness in network security and threat detection is confirmed.

Securing DNS over HTTPS traffic: a real-time analysis tool

24/10/2024

Adversarial attacks in signature verification: a deep learning approach
DOI: https://doi.org/10.11591/csit.v5i3.p215-226

This study develops a handwritten signature verification system using a convolutional neural network (CNN) and hand-crafted adversarial attacks. The model achieved a testing accuracy of 91.35% and a robust validation accuracy of nearly 98%. However, the model's accuracy is reduced to nearly 80% when manually constructed adversarial assaults are introduced. The study emphasizes the need for considering adversarial resilience in deep learning models for classification tasks.

Adversarial attacks in signature verification: a deep learning approach

23/10/2024

Vol 5, No 3 Nov 2024 - https://doi.org/10.11591/csit.v5i3
Computer Science and Information Technologies ISSN 2722-323X, e-ISSN 2722-3221 is an open access, peer-reviewed international journal that publishes original research articles, review papers, and short communications that will have an immediate impact on the ongoing research in all areas of Computer Science/Informatics, Electronics, Communication and Information Technologies.

This journal is recognized (accredited) "SINTA 2" by the Ministry of Education, Culture, Research, and Technology of the Republic of Indonesia (Decree No. 164/E/KPT/2021).

Journal on Computer Science and Information Technologies

20/09/2024

Congratulations to all the authors!!
The Bulletin of Electrical Engineering and Informatics has just released its December 2024 issue (early access).

Alhamdulillah, all published papers in this journal until the October 2024 issue have been indexed by Scopus and can be found at https://www.scopus.com/sourceid/21100826382.

This open-access peer-reviewed journal publishes the results of original research, reviews, surveys, or new data/concepts in all fields of interest, including electrical engineering, electronics, instrumentation, control, robotics, telecommunication, computer engineering, information systems, information technology, and informatics (computer science).

20/09/2024

Congratulations!! The International Journal of Evaluation and Research in Education (IJERE) has just published its December 2024 issue (for early access), and the published papers are available until the October 2024 issue (previous issue) at https://www.scopus.com/sourceid/21100934092.