Jurnal Teknologi Informasi, Komputer, dan Aplikasinya (JTIKA )

PENERAPAN METODE CONVOLUTIONAL NEURAL NETWORK (CNN) UNTUK ANALISIS PRODUKSI BIOGAS DALAM PENGOLAHAN SAMPAH ORGANIK BERBASIS IOT

Endro Ariyanto, M. Hadist, Hilal Hudan Nuha — Wed, 30 Sep 2026 21:38:19 +0800

The development of biogas technology in Indonesia faces challenges in production monitoring and prediction accuracy due to the non-linear complexity of data. This study aims to implement a Convolutional Neural Network (CNN) method integrated with Internet of Things (IoT) technology to analyze and predict biogas production parameters. The system utilizes an ESP32 microcontroller with MQ-4, DHT11, and pH sensors for real-time data acquisition sent to a cloud service. The prediction model was developed using a 1D-CNN architecture trained on 2,400 sensor data points collected at 30-second intervals. Results demonstrate that the CNN model is highly effective in predicting methane gas concentration, achieving an R² of 0.99661 and a MAPE of 1.3888%. Temperature and pH predictions yielded R² values of 0.8982 and 0.6633, respectively, with MAPE remaining below the 10% threshold. Overall, the integration of IoT and CNN provides accurate and stable predictions, serving as an intelligent reference for optimizing biogas production from household organic waste.

SISTEM PEMESANAN LAPANGAN BADMINTON BERBASIS WEB DI LOMBOK TIMUR (STUDI KASUS: ESOCH SPORT)

Lalu Ibnu Agung, Dwi Ratnasari, Royana Afwani — Wed, 30 Sep 2026 21:40:38 +0800

The badminton court booking system in East Lombok still operates manually, requiring customers to visit the location in person. Based on observations, this manual method serves 10–15 customers per day, but only a portion can be confirmed properly. This is due to the use of manual methods that often cause problems such as errors in recording reservations, lost reservation data, and clashing reservation schedules. To overcome these problems, this study aims to design a web-based badminton court booking system to simplify and expedite the booking process. The system development uses the waterfall method, which includes the stages of analysis, design, implementation, testing, and maintenance. The result of this study is a web-based badminton court booking system that is able to process online reservations, display available schedules in real time, and minimize recording errors. Testing using black box testing and User Acceptance Testing (UAT) shows the system runs well, with a user acceptance rate of 86.4% based on responses from 15 respondents. This system improves efficiency and accuracy in the booking process.

KLASIFIKASI CITRA KELAINAN KULIT MENGGUNAKAN RESNET50 DAN MOBILENETV2 PADA DATASET HAM10000

Ida Ayu Dewi Purnama Anjani, Fitri Bimantoro, Halil Akhyar — Wed, 30 Sep 2026 21:42:41 +0800

Penyakit kulit merupakan salah satu masalah kesehatan dengan prevalensi yang cukup tinggi di Indonesia. Proses diagnosis kelainan kulit umumnya dilakukan melalui biopsi atau pemeriksaan visual menggunakan dermoskop yang memerlukan waktu serta ketelitian tinggi. Perkembangan deep learning, khususnya Convolutional Neural Network (CNN), memungkinkan proses klasifikasi citra medis dilakukan secara otomatis dengan tingkat akurasi yang lebih baik melalui pendekatan transfer learning. Penelitian ini bertujuan untuk menganalisis dan membandingkan performa model ResNet50 dan MobileNetV2 dalam klasifikasi tujuh jenis kelainan kulit pada dataset HAM10000. Tahapan penelitian meliputi preprocessing data, augmentasi citra, pembagian dataset menjadi data train, validation, dan test, training model menggunakan bobot pretrained ImageNet, fine-tuning, serta evaluasi menggunakan confusion matrix dan metrik accuracy, precision, recall, dan F1-score. Hasil penelitian menunjukkan bahwa ResNet50 memperoleh performa terbaik dengan accuracy 85,43%, precision 86,26%, recall 85,43%, dan F1-score 85,67%. Sementara itu, MobileNetV2 memperoleh accuracy 83,70%, precision 83,64%, recall 83,70%, dan F1-score 83,63%. Dari sisi efisiensi komputasi, MobileNetV2 memiliki jumlah parameter yang lebih sedikit dan waktu inferensi yang lebih cepat dibandingkan ResNet50. Hasil penelitian menunjukkan bahwa ResNet50 lebih unggul dalam performa klasifikasi, sedangkan MobileNetV2 lebih efisien untuk implementasi pada perangkat dengan sumber daya komputasi terbatas.

PENGEMBANGAN SISTEM INFORMASI KEUANGAN SMP-IT ANAK SOLEH BERBASIS WEB

Muhammad Dzaki Al-Qushoyyi, Nadiyasari Agitha, Moh. Ali Albar — Wed, 30 Sep 2026 21:45:07 +0800

Pengelolaan keuangan sekolah yang masih dilakukan secara manual berpotensi menimbulkan kesalahan pencatatan, keterlambatan penyusunan laporan, serta kesulitan dalam monitoring pembayaran siswa. Berdasarkan hasil observasi dan wawancara dengan pihak sekolah, permasalahan tersebut terjadi pada SMP-IT Anak Sholeh Mataram, di mana proses pembayaran SPP, pencatatan transaksi, dan penyusunan laporan keuangan masih dilakukan menggunakan buku dan spreadsheet sederhana. Penelitian ini bertujuan untuk merancang dan mengembangkan Sistem Informasi Keuangan berbasis web yang mampu membantu pengelolaan administrasi keuangan sekolah secara lebih terstruktur, terintegrasi, dan efisien. Metode yang digunakan dalam penelitian ini adalah Prototyping, yang dilakukan melalui tahapan analisis kebutuhan, perancangan cepat, pembangunan prototype, dan evaluasi user secara iteratif. Sistem yang dikembangkan memiliki fitur pengelolaan data siswa, pembayaran SPP, monitoring tunggakan, laporan keuangan, dashboard monitoring, serta notifikasi pembayaran melalui WhatsApp Gateway. Pengujian sistem dilakukan menggunakan Black Box Testing dan User Acceptance Testing (UAT). Hasil pengujian terhadap prototipe menunjukkan bahwa fungsionalitas sistem telah berhasil mengakomodasi proses pencatatan transaksi keuangan agar lebih terpusat dan terintegrasi, serta menyediakan fitur yang dirancang untuk mempermudah monitoring pembayaran dan penyusunan laporan keuangan secara real-time. Berdasarkan hasil UAT, sistem memperoleh rata-rata persentase penilaian sebesar 92,99% yang berada pada kategori sangat baik. Dengan demikian, Sistem Informasi Keuangan SMP-IT Anak Sholeh Mataram berbasis web dinilai layak digunakan untuk mendukung pengelolaan administrasi keuangan sekolah.

IMPLEMENTASI CONTINUOUS INTEGRATION DAN CONTINUOUS DELIVERY PADA PROYEK APLIKASI CHART LYRIC EDITOR

Ghifary Ahada Azra, Moh. Ali Albar, Sri Endang Anjarwani — Wed, 30 Sep 2026 21:47:20 +0800

Modern software development paradigms demand iterative code integration and delivery to ensure release speed and system stability. However, reliance on manual processes often hinders these standards due to error proneness. In the context of the Chart Lyric Editor application, development stages are still performed manually, leading to human error risks and release artifact inconsistencies. This research aims to address these issues by implementing Continuous Integration (CI) and Continuous Delivery (CD) using the Design Science Research Methodology (DSRM) approach. The solution is designed using GitHub Actions with a separated workflow strategy executed on a Windows environment. This strategy divides the process into two parts: an automated integration workflow (CI) triggered by code changes and a controlled release workflow (CD) via manual triggers. Evaluation was conducted by comparing release cycle efficiency and build consistency between the automated pipeline and the previous manual procedure. The results demonstrate that the CI/CD implementation significantly improves release cycle efficiency, reducing the total process duration from 17 minutes 51 seconds manually to an average of 1 minute 47 seconds. Furthermore, this automation establishes a more structured development workflow, minimizes potential human errors, and maintains the stability and consistency of application release artifacts.

PENERAPAN DEEP LEARNING DENGAN ATTENTION MECHANISM PADA PERAMALAN TRAFIK JARINGAN UNTUK MANAJEMEN QoS ADAPTIF

Muhammad Farhan Zul Fahmi, Bambang Krismono Triwijoyo, Neny Sulistianingsih — Wed, 30 Sep 2026 21:49:59 +0800

Static Quality of Service (QoS) management in government network infrastructures often fails to dynamically adapt bandwidth allocation to fluctuating and non-linear traffic conditions. This study develops an adaptive QoS management system based on a Hybrid Long Short-Term Memory (LSTM)-Attention model for network traffic forecasting at the Communication and Information Technology Office (Diskominfo) of Lombok Barat Regency, serving approximately 200 Regional Government Organization (OPD) users with 700 Mbps dedicated 1:1 ISP capacity. A dataset of 43,488 data points was collected from the WAN interface of a MikroTik Router via RouterOS API during January–May 2026 at 5-minute intervals. The preprocessing pipeline included Z-Score outlier detection with capping, Min-Max Scaling normalization, and Sliding Window transformation (timestep = 60). The model architecture consists of two stacked LSTM layers (128 and 64 units) and one custom Attention Layer, totaling 122,241 trainable parameters, with a sequential 70/15/15 train-validation-test split. Evaluation on the testing set (May 2026) yielded RMSE of 43.66 Mbps and MAE of 27.96 Mbps for download traffic (RX), and RMSE of 30.72 Mbps and MAE of 19.66 Mbps for upload traffic (TX). MAPE values of 19.64% (RX) and 17.95% (TX) fall in the "Good" category. The model outperformed the ARIMA baseline on three of four evaluation metrics, with RMSE improvement of 16.7% and MAE improvement of 28.8%, while MAPE was marginally higher (+1.00 point) than ARIMA, a nuance attributable to the differing sensitivity of scale-dependent versus percentage-based metrics; an ablation study further confirmed the Attention layer's contribution is asymmetric, consistently improving accuracy on upload (TX) traffic and substantially reducing training variance on both signals (62.5-83.1% lower standard deviation across seeds) at a marginal parameter cost (+3.5%). Integration with MikroTik Queue Tree via RouterOS API successfully executed 603 automated adaptive cycles, achieving 26.7% throughput increase, 37.0% delay reduction, 59.3% packet loss reduction, and 49.6% jitter reduction compared to conventional static QoS management.

PENGEMBANGAN APLIKASI SAFE EXAM BROWSER BERBASIS ANDROID TERINTEGRASI MOODLE MENGGUNAKAN METODE RAPID APPLICATION DEVELOPMENT DI SMAN 2 SUKAWATI

Aris Surya Kusuma, A.A. Istri Ita Paramitha, Putri Anugrah Cahya Dewi — Wed, 30 Sep 2026 21:52:13 +0800

The integration of digital technology in education has increased the use of online assessment systems, particularly through Learning Management Systems. However, the implementation of digital exams at SMAN 2 Sukawati still faces challenges related to academic integrity, as students can switch applications, access external resources, and capture exam content during assessments. This study aims to develop an Android-based Safe Exam Browser application integrated with Moodle to create a secure and controlled exam environment. The application utilizes a device locking mechanism that restricts access to other applications and system functions during the exam. The development process applies the Rapid Application Development method, allowing iterative design and rapid adaptation to user requirements. The system includes features such as application locking, screenshot protection, external link blocking, violation detection, and automated penalty mechanisms. Scenario-based functional testing was conducted involving students and teachers to evaluate whether each feature performed as intended. The results indicate that the developed features—including device locking, violation detection, and the penalty mechanism—successfully functioned according to the defined test scenarios. These findings suggest that the proposed system has the potential to support a more secure and controlled digital exam environment, although further evaluation involving user acceptance and comparative testing is needed to confirm its effectiveness in reducing cheating behavior.

SISTEM INFORMASI EKSTRAKURIKULER BERBASIS WEB DI MAN 1 LOMBOK TIMUR MENGGUNAKAN METODE PERSONAL EXTREME PROGRAMMING (PXP)

Husnul Huda Galih Saputra, Dwi Ratnasari, Moh. Ali Albar — Wed, 30 Sep 2026 00:00:00 +0800

Extracurricular management at MAN 1 Lombok Timur currently faces administrative challenges, where 64.7% of attendance and reporting processes are still conducted manually based on survey. This manual system leads to high risks of data loss and inefficiency in monitoring activities. This research aims to design and develop a web-based extracurricular information system using the Personal Extreme Programming (PXP) method to improve administrative efficiency. The PXP methodology consists of requirements, planning, iteration, design, implementation, and testing phases. The developed system integrates key features including digital attendance, gallery management, and automated reporting. Testing was performed using Black Box Testing and User Acceptance Testing (UAT). The results of Black Box testing, covering both positive and negative scenarios, demonstrate that the system is robust and functions according to the requirements. UAT results indicate high user acceptance, suggesting that the system successfully provides a transparent and structured platform for supervisors, administrators, and members. Based on both blackbox and UAT results, this system can replace manual processes in extracurricular management.

SISTEM INFORMASI MONITORING PERKEMBANGAN SISWA SMA KATOLIK KESUMA MATARAM MENGGUNAKAN METODE PROTOTYPING

Ida Ayu Vinaya Anindya, Santi Ika Murpratiwi, Heri Wijayanto — Wed, 30 Sep 2026 21:58:16 +0800

The advancement of information technology drives the education sector toward digital transformation, including in the management of student grades and academic progress reporting. At SMA Katolik Kesuma Mataram, the grade management process is still conducted manually using spreadsheets, leading to data recapitulation errors and delayed academic reporting. This study aims to design and develop a web-based student development monitoring information system to enhance transparency in student academic and non-academic progress management. The system was developed using the prototyping method so that user needs could be accommodated iteratively through feedback from the school. Key features include academic grade input, non-academic recording, student attendance tracking, and visualization with student development curves. The system is also equipped with a rule-base classification mechanism using an IF-THEN structure. Unlike previous studies that apply rule-based approaches without subject differentiation, this study introduces a classification model that distinguishes compulsory and elective subjects based on KKM tolerance ranges, enabling more contextually relevant early identification of student academic conditions. Testing was conducted using white box testing, black box testing, and user acceptance tests to ensure functionality and user acceptance levels. User Acceptance Testing involved 73 respondents consisting of 1 principal, 25 subject teachers, 15 homeroom teachers, 30 parents, and 2 administrators at SMA Katolik Kesuma Mataram, yielding MOS scores of 5,00 for the principal, 4,73 for subject teachers, 4,70 for administrators, 4,30 for parents, and 4,63 for homeroom teachers, indicating that the system has been well accepted by all users.

KLASIFIKASI UNTUK MENDETEKSI PENYAKIT IKAN MENGGUNAKAN MODEL VGG-16 DAN MOBILENETV2

Maftuh Ahnan Al-Kautsar, Fitri Bimantoro, I Gede Pasek Suta Wijaya — Wed, 30 Sep 2026 21:59:59 +0800

Fish disease is a major challenge in the aquaculture sector, causing global economic losses of up to US$6 billion annually. Conventional disease identification relies on visual observation and laboratory testing, which is time-consuming, costly, and requires specialized expertise. This study investigates freshwater fish disease classification using two Convolutional Neural Network (CNN) architectures based on transfer learning, namely VGG-16 and MobileNetV2, combined with sequential hyperparameter tuning. The dataset consists of 2,800 images across eight classes, comprising 2,000 training images and 800 images used for validation and final evaluation. Sequential tuning was performed across five hyperparameters: optimizer, learning rate, dropout, batch size, and number of epochs. The best configuration of VGG-16 (Adam optimizer, learning rate 1e-4, dropout 0.2, batch size 16, and 50 epochs) achieved an accuracy of 99.62%, while MobileNetV2 (Adam, learning rate 1e-3, dropout 0.2, batch size 16, and 50 epochs) achieved an accuracy of 99.37% on the evaluation data. VGG-16 produced a slightly higher accuracy, with a difference of 0.25 percentage points, whereas MobileNetV2 had substantially fewer base-model parameters. These results indicate that both architectures achieved high classification performance under the experimental setting used in this study, while their differences in model complexity may be considered when selecting a model for further deployment. However, the use of the same data for validation and final evaluation represents a limitation that may result in an optimistic estimation of model performance

ANALISIS DESKRIPTIF DAN PREDIKTIF TIME SERIES PADA TRANSAKSI COFFEE SHOP MULTI PRODUK MENGGUNAKAN METODE SARIMA (STUDI KASUS: AMORE ESPRESSO MATARAM)

Wed, 30 Sep 2026 22:01:52 +0800

The rapid growth of the coffee shop industry has encouraged business owners to utilize transaction data as a basis for decision-making. However, transaction data are often underutilized to identify sales patterns and predict future revenue. This study aims to analyze sales patterns and forecast daily revenue at Amore Espresso Mataram using the Seasonal Autoregressive Integrated Moving Average (SARIMA) method. The dataset consists of 18,738 transaction records collected from January to December 2022, including the attributes of Date, Time, Net Sales, and Items. The research consists of data preprocessing, descriptive analysis, stationarity testing using the Augmented Dickey-Fuller (ADF) test, parameter identification through Autocorrelation Function (ACF) and Partial Autocorrelation Function (PACF) plots, SARIMA model development, residual diagnostic checking, model evaluation using Mean Absolute Deviation (MAD) and Mean Absolute Percentage Error (MAPE), and revenue forecasting. The SARIMA model achieved a MAD value of IDR 961,097.80 and a MAPE value of 24.13%, indicating a moderate forecasting accuracy. Descriptive analysis shows that Saturday recorded the highest revenue among the days of the week, while Thursday recorded the lowest revenue. Among the operational periods, the morning period recorded the highest revenue, while the afternoon period recorded the lowest revenue. In addition to generating revenue forecasts, the study provides business insights regarding sales patterns based on operational periods, days of the week, and monthly trends. These findings can support operational planning and data-driven decision-making at Amore Espresso Mataram.

PENGEMBANGAN SISTEM INFORMASI PENCATATAN KEHADIRAN SISWA DAN MONITORING ORANG TUA BERBASIS WEBSITE MENGGUNAKAN PERSONAL EXTREME PROGRAMMING DI MI INTEGRAL BUAH HATI INSANI

Baiq Luthfida Khairunnisa, Royana Afwani, Herliana Rosika — Wed, 30 Sep 2026 22:03:56 +0800

The student attendance recording process at MI Integral Buah Hati Insani Mataram is still carried out manually using attendance books, which increases the risk of data loss, delays in attendance recap generation, and limited access to attendance information for parents. In addition, the school does not yet have an integrated platform for parents to evaluate the school and teachers. This study proposes an integrated web-based student attendance recording and parental engagement system using the Personal Extreme Programming (PXP) methodology through five development iterations. The application was built using Next.js, Prisma ORM, and Neon DB. System testing was conducted using Unit Testing and User Acceptance Testing (UAT). Unit Testing showed that all system functions operated according to user requirements. The UAT results achieved scores of 100% for administrators, 83.71% for teachers, 90.28% for parents, and 100% for the principal, with an overall score of 90.34%, categorized as “Highly Feasible.” The developed system integrates student attendance recording, parental monitoring, communication, and school and teacher evaluation within a single web-based platform.

SISTEM INFORMASI PERHITUNGAN PAJAK PENGHASILAN PASAL 21 BERBASIS WEB (STUDI KASUS PADA KANTOR KONSULTAN PAJAK BUDI SATRIYA)

Putu Risanti Iswardani, I Gusti Ayu Agung Mas Aristamy — Wed, 30 Sep 2026 22:05:35 +0800

The computation of Income Tax (PPh) Article 21 utilizing the gross-up method at the Budi Satriya Tax Consultant Office currently relies on user-defined formulas within Microsoft Excel. This conventional mechanism exhibits limitations in time efficiency and is susceptible to formula-related human errors. This study aims to design and implement a web-based PPh Article 21 calculation information system to enhance data processing accuracy and efficiency. Data collection techniques encompassed observation, interviews, documentation, and literature review. The system development adopted the Waterfall model, implemented using PHP 8 and Visual Studio Code. System evaluation was conducted through functional verification via Black-box Testing and User Experience (UX) analysis. The results indicate that the developed information system functions correctly and fulfills user requirements. The UX analysis confirms that the system provides significant simplification in the tax calculation process, reduces the potential for calculation errors, and optimizes processing time efficiency compared to the predecessor system.

ESTIMASI KONSENTRASI PERLAKUAN LARUTAN NUTRISI HIDROPONIK PAKCOY BERBASIS CITRA MENGGUNAKAN YOLOV8S DAN EFFICIENTNETV2-S

Aniza Wulandari Wulandari, Nike Dwi Grevika Drantantiyas, Ahmad Suaif Suaif — Wed, 30 Sep 2026 22:07:16 +0800

Nutrient availability is important in hydroponic pakcoy cultivation, but image-based monitoring requires a clearly defined target and a reliable estimation pipeline. This study develops and evaluates a cascaded computer-vision method using YOLOv8s to localize pakcoy leaves and EfficientNetV2-S to estimate the experimentally assigned nutrient-solution treatment concentration from leaf regions. After dataset auditing, 1,995 unique annotated images were divided into 1,595 training, 200 validation, and 200 independent test images covering 250, 500, 750, and 1,000 ppm. YOLOv8s was trained for 50 epochs with 320 × 320-pixel inputs, while EfficientNetV2-S used 224 × 224-pixel regions of interest, AdamW optimization, and Huber loss. On the independent test set, YOLOv8s achieved precision of 0.9173, recall of 0.8870, mean average precision at 0.50 intersection-over-union of 0.9598, and mean average precision at 0.50–0.95 of 0.8428, with usable leaf regions obtained from 170 of 200 images. Conditional on successful localization, EfficientNetV2-S achieved a mean absolute error of 42.76 ppm, root mean square error of 107.37 ppm, mean absolute percentage error of 5.71%, and R-squared of 0.8577. The results support detector-guided estimation within the four observed treatment levels, while generalization to unseen concentrations and cultivation conditions remains unvalidated.

ND-AOMDV: SELEKSI PENERUSAN RREQ BERBASIS KEPADATAN NODE TETANGGA PADA PROTOKOL ROUTING AOMDV DI JARINGAN MANET BERKEPADATAN TINGGI

Andy Hidayat Jatmika, Raphael Bianco Huwae — Wed, 30 Sep 2026 22:08:52 +0800

Abstract

Route discovery in the Ad hoc On-demand Multipath Distance Vector (AOMDV) protocol relies on flooding Route Request (RREQ) packets, a mechanism that becomes increasingly wasteful as more nodes share the same radio neighborhood. In dense Mobile Ad Hoc Networks (MANETs), most rebroadcasts carry redundant information and only add channel contention, collisions, and control traffic. This paper presents Node Density-AOMDV (ND-AOMDV), in which every intermediate node regulates its own participation in route discovery using only its number of active neighbors. Nodes with at most k=11 neighbors always rebroadcast, whereas nodes in crowded neighborhoods rebroadcast with a probability inversely proportional to their neighbor count. The scheme was implemented in NS-2.35 and compared with standard AOMDV in a 1000 × 1000 m² area with 25 to 275 nodes. Standard AOMDV remains superior up to 175 nodes, but from 200 nodes onward ND-AOMDV delivers higher throughput and packet delivery ratio (PDR) with lower end-to-end delay and routing overhead. With 275 nodes, ND-AOMDV sustains a PDR of 92.37% versus 82.9% and lowers delay and routing overhead by 13.2% and 13.0%, respectively. Neighbor density is therefore a lightweight and effective criterion for controlling RREQ flooding in dense MANETs.