Article Per Year (5 Year)
p-Index From 2021 - 2026
0.61
P-Index
This Author published in this journals
All Journal Abdimas Talenta : Jurnal Pengabdian Kepada Masyarakat KOMPUTEK
Machfudz, Emir Syarif
Unknown Affiliation
Agriculture, Biological Sciences & Forestry Civil Engineering, Building, Construction & Architecture Computer Science & IT Control & Systems Engineering Electrical & Electronics Engineering Law, Crime, Criminology & Criminal Justice Mechanical Engineering Social Sciences

Published : 3 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 3 Documents
Search

 1 
Pengaruh Struktur Hidden Layer pada Model Multilayer Perceptron terhadap Prediksi Status Akademik Mahasiswa Machfudz, Emir Syarif; Putra, Farhansyah; Alfi, Khairuman; Armansyah, Armansyah
KOMPUTEK Vol. 9 No. 1 (2025): April
Publisher : Universitas Muhammadiyah Ponorogo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24269/jkt.v9i1.3094

Abstract

Fenomena mahasiswa yang berisiko dropout menjadi perhatian penting bagi institusi pendidikan tinggi karena berdampak pada kualitas dan akreditasi lembaga. Penelitian ini bertujuan memanfaatkan jaringan syaraf tiruan (JST) dengan arsitektur Multilayer Perceptron (MLP) untuk memprediksi status akademik mahasiswa berdasarkan data historis. Analisis difokuskan pada pengaruh jumlah hidden layer terhadap kinerja model, dengan membandingkan konfigurasi 3 hidden layer dan 4 hidden layer. Dataset yang digunakan berisi 1.337 data mahasiswa, mencakup fitur seperti tahun masuk, lama studi, IPK, total SKS, dan biaya kuliah. Data diproses melalui normalisasi MinMaxScaler dan diolah menggunakan Scikit-learn dengan fungsi aktivasi ReLU, optimizer Adam, dan fungsi loss Log Loss. Hasil menunjukkan bahwa konfigurasi 3 hidden layer menghasilkan akurasi tertinggi sebesar 95,52% dengan waktu pelatihan yang lebih cepat dibandingkan konfigurasi 4 hidden layer yang hanya mencapai akurasi 94,03%. Temuan ini menunjukkan bahwa jumlah hidden layer yang optimal dapat meningkatkan performa model sekaligus menjaga efisiensi komputasi. Model yang dikembangkan memiliki potensi untuk membantu institusi pendidikan dalam mencegah dropout melalui deteksi dini dan intervensi yang tepat sasaran. 
ABA DOA (Android-Based Attendance for D-III Accounting Lecturers) Using the USU Vocational Faculty Location Authorization Feature Machfudz, Emir Syarif; Sinaga, Annisa Fitri; Zufria, Ilka; Syarif, Firman
ABDIMAS TALENTA: Jurnal Pengabdian Kepada Masyarakat Vol. 9 No. 1 (2024): ABDIMAS TALENTA: Jurnal Pengabdian Kepada Masyarakat
Publisher : Talenta Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32734/abdimastalenta.v9i1.15457

Abstract

Abstract This research details the development of Android-based Attendance Application for D-III Accounting Lecturers (ABA DOA) using Appsheet platform and utilizing LATLONG() and DISTANCE() functions for location validation. The development method applied is Rapid Application Development (RAD). ABA DOA provides convenience for lecturers with a self-photo feature as proof of validation, while integrating GPS and Google Maps to validate the location of attendance. The application results show efficiency in academic management with an optimal level of security. The LATLONG() function is used to determine the location coordinate point, while DISTANCE() is used to measure the distance between the authorized coordinate point and the attendance location. This research highlights the importance of technological innovation in supporting academic activities and campus management. The conclusion of this research is that ABA DOA successfully optimizes the attendance process with a high level of validation and security, making a positive contribution to the efficiency and effectiveness of study program management.
Implementasi Kriptografi AES pada Sistem Enkripsi Gambar berbasis Web dengan Kunci Ephemeral yang dihasilkan API Machfudz, Emir Syarif; Suhardi
KOMPUTEK Vol. 9 No. 2 (2025): Oktober
Publisher : Universitas Muhammadiyah Ponorogo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24269/jkt.v9i2.3469

Abstract

This study implements the Advanced Encryption Standard (AES) algorithm in a web-based image encryption system utilizing an ephemeral key generated by an API. The research is motivated by the increasing threats to digital data security, such as identity theft and information manipulation, as well as the weaknesses of static key usage in AES encryption. The developed system consists of a client-side web application and a RESTful API server. The client side, built with HTML, CSS, and JavaScript, receives image input, requests the API for a key, and performs AES-256 encryption in Cipher Block Chaining (CBC) mode with a unique Initialization Vector (IV). The server side generates a random 256-bit key, stores it temporarily for approximately 30 seconds, and sends the key ID to the client. Testing was conducted on images sized 50 KB, 200 KB, and 1000 KB to measure encryption time. The results show that encryption time increases linearly with image size, with minimal overhead that does not affect performance. The ephemeral key approach enhances security by limiting the opportunity for unauthorized access to the key, while client-side encryption ensures that raw image data is never transmitted to the server. These findings demonstrate that implementing AES with ephemeral keys effectively improves the security of web-based image encryption.
Co-Authors Alfi, Khairuman Armansyah Armansyah Firman Syarif Ilka Zufria Putra, Farhansyah Sinaga, Annisa Fitri Suhardi