Jurnal Teknologi Informasi dan Terapan

Mobile App for Incubator Monitoring to Optimize Quail Egg Production

Agus Nur Khomarudin — 2026-06-29

Incubation is a crucial process in egg hatching, where eggs must be maintained under optimal temperature and humidity conditions. To ensure the stability of these parameters, a monitoring system is needed one that operates continuously, provides sufficient accuracy, and is easily accessible to users. This study aims to design a Mobile Application for Monitoring Temperature and Humidity in Quail Egg Incubators that is both valid and effective, as a means to enhance and optimize quail egg production. The research follows the Agile Method, consisting of the following phases: Plan, Design, Develop, Test, Deploy, Review, and Launch. The Mobile App for Incubator Monitoring to Optimize Quail Egg Production underwent product testing, including a validity test conducted by several experts in computer and mobile applications. The test resulted in a score of 0.78, which falls into the valid category. The effectiveness test yielded a score of 0.35, indicating a moderate level of effectiveness. The implementation of this monitoring application has shown a significant positive impact on operational efficiency in quail farming, particularly at Nisya Farm. User experience was evaluated through direct observation and interviews with respondents. Overall, users reported that the app is easy to use, thanks to its simple and intuitive interface. User feedback strengthens the argument that technological integration not only brings technical benefits but also boosts farmers’ confidence in managing the incubation process. This is essential, as the success of technology adoption heavily depends on how comfortable and user-friendly it is for its intended users.

Comparative Analysis of AES-GCM and ChaCha20-Poly1305 in IoT Data Encryption Based on ESP32

Dwi Dinda Meylani Angelina — 2026-06-29

Internet of Things (IoT) devices have become integral to modern applications, yet their resource constraints pose significant challenges for implementing robust security measures. While AES-GCM and ChaCha20-Poly1305 are widely recognized Authenticated Encryption with Associated Data algorithms, their comparative performance on resource-limited microcontrollers like ESP32 remains underexplored, particularly regarding execution time, memory usage, power consumption, and throughput. This research aims to conduct a comprehensive performance analysis of both algorithms for securing IoT data transmission on ESP32-WROOM-32D microcontrollers. The study implements both algorithms using Arduino IDE 2.3.6, leveraging ESP32's hardware acceleration for AES-GCM and optimized software implementation for ChaCha20-Poly1305. Performance evaluation encompasses various payload sizes (16, 64, and 256 bytes) with precise measurements of execution time using micros() function, memory usage via ESP.getFreeHeap(), and power consumption through shunt resistor analysis. The results reveal distinct performance characteristics between the two algorithms across all evaluated metrics, providing valuable comparative insights for IoT developers to select optimal cryptographic solutions based on specific application requirements and resource constraints, thereby enhancing security implementation on embedded systems.

Management of Potential Mental Health and Behavioral Disorders for College Students Using Integrated Applications: Implementation of Human-Centered Design

Rinda Nurul Karimah — 2026-07-03

Symptoms of mental disorders among college students have risen recently. A comprehensive synthetic review using a socio-ecological model as a guiding framework reveals that college students' mental health is influenced by several dynamic and interconnected factors at the individual, interpersonal, institutional, community, and policy levels, all of which can contribute to stress, anxiety, and depression. Several information and communication technology (ICT) services have been developed to improve healthcare, including the Depression, Anxiety, and Stress Scale 21 (DASS21) instrument. However, these applications do not receive continual monitoring and support from mental health professionals such as psychologists. To address the mental health difficulties of adolescent college students, it is critical that the services established provide appropriate interventions and successfully identify, detect, and address student mental health concerns. As a result, when creating an application interface, a Human-Centered Design (HCD) approach is required, which prioritizes human interaction to provide a more intuitive, precise, and user-friendly user experience. The success of the HappyMind app design was demonstrated by testing it on target users, namely college students, and end users, especially psychologists who served as evaluators. The results demonstrate that the HappyMind application design achieved an average score of 4 or 5, particularly for simplicity of use, text clarity, comfort, and visual appeal.

Design of a Naive Bayes–Based Adaptive Modulation Model in a Time-Varying Channel Environment

Rosabella Ika Yuanita — 2026-07-03

Orthogonal Frequency Division Multiplexing (OFDM) systems require an effective modulation adaptation mechanism to maintain transmission reliability over dynamic and noise-affected channels. This study proposes a machine learning–based adaptive modulation method using Naive Bayes classification to select the most appropriate modulation scheme—BPSK, QPSK, or 16-QAM—based on Signal-to-Noise Ratio (SNR) values. The Naive Bayes model is trained using the probabilistic performance distributions of each modulation scheme, enabling optimal modulation mode prediction under various channel conditions. Simulation results demonstrate that the proposed adaptive method achieves a lower Bit Error Rate (BER) compared to fixed modulation schemes, particularly under low to medium SNR conditions. Furthermore, the Naive Bayes–based approach exhibits more stable performance, especially in recovering transmitted messages. BER curves and demodulated message results indicate that the artificial intelligence–based adaptive scheme using Naive Bayes improves the reliability of transmitting the text message “HELLO WORLD” across an SNR range of –5 dB to 15 dB. These findings confirm that integrating intelligent methods into adaptive OFDM modulation provides an effective solution for wireless communication in fluctuating channel environments.

Culturally Adaptive AI System for Wayang Character Visualization and Recognition for Children

Yoga Rarasto Putra — 2026-06-29

This study developed an artificial intelligence–based drawing and classification system to support the visual reinterpretation of wayang characters for children while preserving their cultural identity. The research addressed the declining interest of younger generations in traditional cultural heritage by introducing a visually engaging and culturally adaptive digital approach. A generative model based on StyleGAN-3 was trained to produce child-friendly visual adaptations of ten wayang characters, while a ResNet-18 classification model was implemented to recognize character images. The dataset consisted of 400 training images and 60 testing images, including children’s drawings used to evaluate model generalization. Image preprocessing and data augmentation techniques were applied to improve model robustness. The classification model achieved an overall accuracy of 87%, indicating strong capability in recognizing distinctive visual characteristics of wayang characters across varied visual styles. In addition, a visual preference evaluation involving children showed that several generated characters received positive responses, particularly those with balanced proportions and expressive features. The results demonstrated that the proposed system can function as an interactive cultural learning medium and provide an innovative strategy for introducing traditional wayang characters to digital-native children.

Development of a Machine Learning Cumulative GPA Prediction Model using Explainable AI

Fathinah Izzati — 2026-06-29

This study aims to develop an accurate, transparent, and interpretable model for predicting students’ Cumulative Grade Point Average (GPA) using an Educational Data Mining approach. The study adopts the Knowledge Discovery in Databases (KDD) framework, which includes data preprocessing, Z-transformation normalization, and feature selection. Three machine learning algorithms, namely Random Forest, XGBoost, and Support Vector Machine (SVM), are compared to determine the best-performing model. Model evaluation is conducted using a 10-fold cross-validation scheme with Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) metrics to ensure generalization capability. To address the black-box nature of machine learning models, Explainable Artificial Intelligence (XAI) techniques are applied using SHAP and LIME to provide both global and local interpretability of the predictions. The results indicate that XGBoost Regression achieves the best performance with the lowest error values. Previous GPA, attendance rate, and study duration are identified as the most influential predictors. The integration of XAI enables deeper insights for educators in supporting data-driven decision-making. Therefore, the proposed model has strong potential to be implemented as an early warning system for more effective and measurable academic interventions.

Machine Learning–Based Recommendation System for Optical Distribution Point Placement in Fiber Access Networks

Widiatry Widiatry — 2026-06-29

The rapid expansion of the digital economy requires reliable telecommunication infrastructure, particularly fiber optic access networks that provide high-speed broadband connectivity. One critical component in these networks is the Optical Distribution Point, which functions as a distribution node connecting optical infrastructure to end users. However, ODP placement is often determined manually, leading to inefficient resource utilization and inconsistent decision-making. This study aims to develop a data-driven recommendation system for optimal ODP placement. The proposed approach integrates spatial feature engineering with supervised machine learning techniques to analyze infrastructure capacity, spatial distance, and customer distribution. Several algorithms were evaluated, including Random Forest, Logistic Regression, K-Nearest Neighbors, Gradient Boosting, and a Stacking Ensemble model, while Synthetic Minority Oversampling Technique was applied to address class imbalance. Model performance was evaluated using Precision, Recall, F1-score, ROC-AUC, and Normalized Discounted Cumulative Gain. The results show that Gradient Boosting achieved the highest performance with an F1-score of 0.8986 and ROC-AUC of 0.96, while the Stacking Ensemble model demonstrated stable ranking performance with a mean NDCG of 91.75%. The proposed system improves the efficiency and accuracy of ODP placement planning and supports data-driven telecommunication infrastructure development.

Multivariate LSTM with SLO-Aware Loss for Virtual Machine Workload Prediction on Cloud Data Center

Agus Hariyanto — 2026-06-29

Accurate virtual machine (VM) workload prediction is a key component of cloud resource management, particularly to support auto-scaling and to maintain Service Level Objectives (SLOs). In conventional prediction models that rely on symmetric loss functions such as Mean Squared Error (MSE), under-prediction errors are treated equivalently to over-prediction errors, even though under-prediction carries significantly more severe operational consequences — it directly triggers capacity shortages and SLO violations. This study proposes a CPU workload prediction approach based on a multivariate Long Short-Term Memory (LSTM) network enhanced with an SLO-aware loss, an asymmetric loss function that penalizes under-prediction ten times more heavily than over-prediction. Experiments are conducted on a subset of 25,000 rows from the Bitbrain GWA-T-12 fastStorage dataset with four input features (CPU, memory, network received, network transmitted), using a fixed random seed for reproducibility. Two models are trained and compared: one with SLO-aware loss and one with standard MSE as baseline, both sharing identical architecture and hyperparameters. The primary evaluation metric is the under-prediction rate, which directly quantifies SLO violation risk. Results show that the SLO-aware model achieves an under-prediction rate of 0.04%, compared to 0.16% for the MSE baseline — a fourfold reduction. These findings empirically confirm that SLO-aware loss effectively directs the model toward conservative predictions that protect SLO compliance, establishing loss function design as a critical and actionable dimension in cloud VM workload prediction.

A Chatbot-Based Expert System for Food Crop Disease Diagnosis

Saniyatul Mawaddah — 2026-06-29

The agricultural sector is very important for Indonesia's food security, however, plant diseases pose a serious threat that can significantly reduce crop yields. The limited availability of agricultural experts often hinders farmers from obtaining rapid and accurate diagnoses. To address this issue, this research develops a chatbot-based expert system using the forward chaining method to assist farmers in conducting self-diagnosis. This method works by drawing conclusions from the symptoms entered, while the chatbot provides real-time interactions that are easy to use. System testing shows good performance, black box testing ensures that all features operate without errors, with a diagnostic accuracy of 87.5 percent, as 42 out of 48 cases correspond with expert assessments. Furthermore, usability testing with 52 respondents yields a System Usability Scale score of 79.18, categorized as good. The results of this research indicate that the developed system is accurate, efficient, and practical, with the potential to serve as a widely applicable solution to help

Web-Based Smart Aquaculture: Comparative Analysis of Mamdani, Sugeno, and Tsukamoto Fuzzy Inference Systems for Shrimp Pond Water Quality Assessment

Santi santi — 2026-06-29

Indonesia possesses vast marine and aquaculture potential; however, national shrimp production in 2024 achieved only 56.67% of its target, largely due to suboptimal water quality management. To address this issue, an intelligent classification system capable of handling uncertainty in aquaculture environments is required. This study presents a comparative evaluation of three fuzzy inference systems (FIS), namely Mamdani, Sugeno, and Tsukamoto, for shrimp pond water quality classification based on four key parameters: temperature, pH, salinity, and dissolved oxygen (DO). Water quality conditions were categorized into four classes: Good, Medium, Bad, and Very Bad using trapezoidal membership functions and expert-defined reference labels derived from aquaculture water quality standards. The dataset consisted of 994 water quality records collected from shrimp ponds in Surabaya, Indonesia, during the period from December 2024 to April 2025. Experimental results indicate that the Mamdani method produced the highest consistency with the expert-defined reference rules, achieving an agreement accuracy of 0.800, precision of 0.825, recall of 0.800, and F1-score of 0.797. In comparison, both Sugeno and Tsukamoto produced lower performance with an accuracy of 0.700 and F1-score of 0.728, although they achieved slightly higher precision values of 0.880. The findings indicate that the Mamdani fuzzy inference system provides more stable and consistent inference behavior relative to the predefined aquaculture reference rules for shrimp pond water quality assessment. Furthermore, the proposed web-based monitoring system demonstrates the practical potential of fuzzy logic approaches in supporting sustainable smart aquaculture management and environmental monitoring.

Deep Learning-Based Tomato Leaf Disease Classification Using CNN, EfficientNetB0, and InceptionResNetV2

Rifqi Aji Widarso — 2026-06-29

Tomato leaf diseases threaten agricultural productivity because symptoms such as early blight, late blight, leaf mold, septoria leaf spot, and yellow curl virus often produce visually similar color changes, necrotic lesions, and leaf deformation. Manual visual diagnosis is subjective and depends heavily on expert experience; therefore, image-based deep learning is a relevant approach for supporting preliminary disease identification. This study compares five deep learning architectures, namely a custom convolutional neural network, EfficientNetB0, MobileNetV2, DenseNet121, and InceptionResNetV2, for classifying six tomato leaf categories using 7,192 images from a Kaggle dataset. The research workflow includes dataset preparation, image resizing and normalization, model training using the Adam optimizer, and evaluation through accuracy, loss, precision, recall, F1-score, and confusion matrix analysis. Based on the notebook results, EfficientNetB0 achieved the best validation accuracy of 89.44% after 20 epochs, followed by MobileNetV2 at 85.12%, DenseNet121 at 82.77%, the custom CNN at 70.69% test accuracy, and InceptionResNetV2 at 45.76% test accuracy. The results indicate that lightweight transfer learning models are more suitable for medium-sized agricultural image datasets than large architectures trained for only a few epochs. Future work should validate the model using real field images, harmonize all models on the same test set, and report class-wise metrics to ensure reliability before deployment as a farmer-oriented diagnostic support system.

LSTM Approached for Cassava Tapai Ripeness Identification

Shabrina Choirunnisa — 2026-06-29

Tapai singkong (cassava tapai) is a traditional Indonesian fermented food product whose quality is highly dependent on precise control of the fermentation process. Inconsistent fermentation outcomes arise from fluctuating environmental conditions including temperature, humidity, and fermentation gas levels making it difficult to reliably determine ripeness status without objective measurement tools. This study addresses the challenge of automated ripeness prediction by providing a controlled, head-to-head comparison of four machine learning approaches Logistic Regression, Support Vector Machine (SVM), Random Forest, and LSTM-based Recurrent Neural Network (RNN) on a single, uniformly preprocessed dataset of 600 time-series observations across three ripeness classes (unripe, ripe, overripe), collected from 10 fermentation trials spanning 60 hours each. All models were evaluated under identical preprocessing and hyperparameter settings using accuracy, precision, recall, F1-score, and confusion matrices to reveal per-class behavior. LSTM yielded the best test performance (96.46% accuracy; macro F1 = 0.93), Random Forest followed closely (93.70% accuracy; macro F1 = 0.94), while SVM and Logistic Regression obtained 91.28% and 90.31% accuracy respectively. This paper discusses the trade-off between predictive performance, temporal modeling capability, and interpretability, and recommends LSTM for high-accuracy quality control deployments where temporal dependencies are critical, and Random Forest as a strong, interpretable alternative for resource-constrained environments. Per-class metrics and experimental artifacts are provided to support reproducibility and practical adoption in traditional food production monitoring.

An Intelligent IoT-Enabled Vermiculture Monitoring and Control System Based on Fuzzy Inference Approach (Case Studi : Siscamling)

Nur Hayati Mufarrihah — 2026-07-03

Cultivating earthworms necessitates meticulous environmental control to ensure optimal growth, reproduction, and substrate integrity. Conventional vermiculture approaches depend significantly on manual monitoring, resulting in irregular moisture and pH management, hence diminishing production and resource efficiency. This research presents an intelligent Internet of Things (IoT)-facilitated monitoring and control system utilising a Sugeno fuzzy inference method for automated environmental management in vermiculture. The system amalgamates sensors for soil moisture, temperature, soil pH, water pH, and total dissolved solids with a microcontroller and cloud platform to facilitate real-time monitoring and remote oversight. Fuzzy logic is utilised to manage environmental uncertainty and ascertain suitable control measures, including irrigation, fertiliser application, and ventilation. Experimental findings indicate that the proposed system proficiently sustains substrate moisture within the ideal range of 15–30% and regulates pH values between 6.0 and 7.2. Automated reactions were effectively initiated under diverse environmental conditions, including irrigation activation at a 40% moisture threshold and nutrient correction in mildly acidic settings. The system attained dependable sensor functionality with negligible transmission latency and enhanced resource efficiency by minimising excessive irrigation and substrate waste. The amalgamation of IoT with fuzzy control offers a scalable, adaptive, and sustainable approach to smart vermiculture management.

A Multi Criteria Framework for Ice Block Production Systems Integrating Machine Performance and Floating Photovoltaics

A.A Chrisna Firman Alamsyah Putra — 2026-07-03

The increasing demand for ice blocks in small island fisheries requires an efficient and sustainable production system that accounts for both technical and energy constraints. This study aims to identify the optimal ice block production system for Bungin Island by integrating machine characteristics, ice demand, and solar energy potential. Ice demand is estimated from fish catch data using a ratio-based approach, while technical evaluation includes criteria such as production capacity, unit weight, quantity, harvesting time, power consumption, specific energy consumption (SEC), and cost.

A multi-criteria decision analysis (MCDA) framework combining the CRITIC and TOPSIS methods is applied to determine criteria weights and rank alternatives. The results indicate that harvesting time and energy efficiency (SEC) are the most influential factors in decision-making. Among the evaluated alternatives, machines with lower SEC and shorter production cycles demonstrate superior performance. Furthermore, the analysis confirms that integrating photovoltaic systems is feasible and can support energy requirements under spatial constraints.

This study provides a systematic framework for optimizing ice block production systems, contributing to sustainable fisheries infrastructure development in small island regions.