cover
Contact Name
Jumanto
Contact Email
jumanto@mail.unnes.ac.id
Phone
+628164243462
Journal Mail Official
sji@mail.unnes.ac.id
Editorial Address
Ruang 114 Gedung D2 Lamtai 1, Jurusan Ilmu Komputer Universitas Negeri Semarang, Indonesia
Location
Kota semarang,
Jawa tengah
INDONESIA
Scientific Journal of Informatics
ISSN : 24077658     EISSN : 24600040     DOI : https://doi.org/10.15294/sji.vxxix.xxxx
Scientific Journal of Informatics (p-ISSN 2407-7658 | e-ISSN 2460-0040) published by the Department of Computer Science, Universitas Negeri Semarang, a scientific journal of Information Systems and Information Technology which includes scholarly writings on pure research and applied research in the field of information systems and information technology as well as a review-general review of the development of the theory, methods, and related applied sciences. The SJI publishes 4 issues in a calendar year (February, May, August, November).
Articles 131 Documents
In Vivo Diagnostic Automation: Identification of Malaria Parasites from Red Blood Cells Using Image Segmentation and Convolutional Neural Network Methods Huda, Nurul; Prihandoko, Prihandoko; Dewi, Alfa Yuliana
Scientific Journal of Informatics Vol. 12 No. 2: May 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i2.23502

Abstract

Purpose: This study aims to address the limitations of conventional malaria diagnosis—namely, its reliance on manual microscopy, which is time-consuming, labor-intensive, and prone to human error—by developing an automated diagnostic system using the Inception V3 convolutional neural network. The focus is on accurately identifying the four main Plasmodium species responsible for malaria (P. falciparum, P. malariae, P. ovale, and P. vivax) through image-based analysis of red blood cells. The study’s significance lies in its contribution to scalable, AI-assisted diagnostic solutions that support national and global malaria elimination goals, particularly in high-burden countries such as Indonesia. Methods: This study utilized an experimental approach based on a dataset of 194 microscopic images of red blood cells, each labeled according to one of four Plasmodium species. The process involved image enhancement through pre-processing techniques—illumination correction, contrast adjustment, and noise filtering—followed by segmentation using the Otsu thresholding method to isolate parasite-infected cells. Two classification models were applied: Inception V3, a deep learning convolutional neural network, and a traditional Support Vector Machine (SVM), with both evaluated for their accuracy in species identification. Result: The findings revealed that the Inception V3 model significantly outperformed the Support Vector Machine (SVM), achieving highest accuracy of 100%, at select epochs and an average accuracy of 97.93%, with 98.32% validation accuracy compared to 82% for SVM. The high performance of Inception V3 is attributed to its deep architecture, consisting of over 23 million parameters, which enables superior feature extraction and classification of Plasmodium species. These results confirm that CNN-based models, particularly Inception V3, are more effective than traditional machine learning approaches for automated malaria diagnosis. Novelty: In identifying four species of Plasmodium, this study presents a very simple yet highly accurate technique using an Inception V3 model. The method represents 100% accuracy in its multi-class detection as opposed to earlier works concentrating on binary classifications. It therefore adds real usefulness in high-burden, low-resource settings such as Indonesia through working on the improvement of diagnosis and on speedier detection of malaria.
Rice Price Forecasting for All Provinces in Indonesia Using The Time Series Clustering Approach and Ensemble Empirical Mode Decomposition Ilmani, Erdanisa Aghnia; Sumertajaya, I Made; Fitrianto, Anwar
Scientific Journal of Informatics Vol. 12 No. 1: February 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i1.23536

Abstract

Purpose: Accurate forecasting of rice prices is essential to ensure food security and a healthy economy for a country like Indonesia. Problems regarding time-series phenomena, such as trends or seasonality, are problematic for traditional approaches like ARIMA (Autoregressive Integrated Moving Average). This study analyzes the effect of EEMD (Ensemble Empirical Mode Decomposition) combined with time-series data clustering on forecasting accuracy. Methods: From 2009 until 2023, the thirty-two Indonesian provincial rice prices were grouped monthly into time-series clusters using hierarchical clustering, average linkage, and DTW (Dynamic Time Warping). After clusterization, the time series were decomposed using the ensemble EEMD method to extract their IMFs (Intrinsic Mode Functions) and residual components. Each IMF was assigned an ARIMA model. The model forecast was generated by adding all individual estimates. MAPE (Mean Absolute Percentage Error) was used to measure the model's performance. Result: The prices were divided into three clusters with an optimized region. Price changes are well captured through EEMD, where the residual components contributed predominantly to the long-term trends. The validation of the prediction showed MAPE values under 10% for the majority of the provinces, which indicates a relatively accurate prediction. On the other hand, some regions had inaccuracies that were higher than others due to uncontrollable fluctuations. Novelty: This study integrates clustering with EEMD decomposition for monthly rice price forecasting using data from 32 Indonesian provinces from 2009 - 2023, offering a novel approach that improves traditional techniques. The model can capture distinct regional price patterns and provide essential information to policymakers to manage rice supply and price stabilization. Further studies can develop external hybrid models with economic variables.
Analysis of Service Package Improvement Factors by Users of Fingerprint Analysis Applications Using TAM and EUCS Methods Kusuma, Teja; Widodo, Aris Puji; Nugraheni, Dinar Mutiara Kusumo
Scientific Journal of Informatics Vol. 12 No. 1: February 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i1.23779

Abstract

Purpose: The research aims to determine the factors that encourage users to upgrade service packages, provide actual recommendations for application developers to improve service quality, and provide insights for the development of similar applications in the future. Future Link is an application that offers efficient online fingerprint analysis with the support of artificial intelligence (AI). Despite these advantages, user acceptance and satisfaction remain challenges that hinder service package upgrades. Methods: This study combines the TAM and EUCS models to identify the variables that influence users' intentions to upgrade their service packages in a fingerprint analysis application. TAM includes key constructs such as perceived usefulness and attitude toward using, while EUCS encompasses dimensions like content, accuracy, format, timeliness, and ease of use, which collectively represent users’ satisfaction with the quality of information provided by the application. Additionally, two external variables curiosity and buying intention were added. A quantitative approach was employed using a survey method, with data collected from 151 respondents who are active users of the application. The data were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM) through Smart-PLS 3.29. The analysis process included instrument development based on validated indicators, testing for construct reliability and validity, followed by evaluation of the structural model to assess the significance and explanatory power of the hypothesized relationships among variables. Result: The findings show that eight out of nine hypotheses were accepted. Significant variables influencing attitude toward use include curiosity, content, format, accuracy, perceived usefulness, and ease of use. User attitude significantly influences satisfaction, which in turn impacts service package upgrade intentions. Novelty: The combination of two theoretical models the TAM and EUCS model is used to analyze user behavior in the context of service package upgrades in fingerprint analysis applications. While previous studies have applied TAM and EUCS separately to evaluate general technology adoption or information satisfaction, this study integrates both models to provide a more comprehensive framework that considers both perceptual and experiential factors influencing user decisions. Furthermore, this study introduces two additional variables curiosity and purchase intention to reflect emerging user motivations in the use of digital services. The results are expected to support the development and improvement of applications in the future.
Information System Evaluation Framework to Improve Teacher and Education Personnel Competency (GTK Room): Extended Hot-Fit Framework Approach Waluyo, Retno; Hariguna, Taqwa; Setiawan, Ito
Scientific Journal of Informatics Vol. 12 No. 2: May 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i2.23841

Abstract

Purpose: This study aims to identify the factors that influence users in the implementation of the GTK (Teachers and Educational Staff) Room System among elementary school teachers in Banyumas Regency, Central Java, Indonesia. Methods: This study employed the HOT-Fit (Human, Organization, and Technology Fit) Framework approach, with the addition of the 'Behavioral Intention to Use' variable on the Human dimension and the 'Organizational Culture' variable on the Organizational dimension. The sample consisted of 147 elementary school teachers from Banyumas Regency, Central Java, Indonesia. Data were analyzed using SmartPLS to identify the variables that influence user behavior. Result: The results of this study indicate that certain relationships between variables do not have a significant influence on others. Specifically, User Satisfaction and Behavioral Intention to Use do not significantly affect Net Benefit. Additionally, Information Quality does not have a significant effect on System Use. Furthermore, System Quality does not significantly influence User Satisfaction or Behavioral Intention to Use. Meanwhile, other variable relationships were found to significantly impact the successful implementation of the GTK (Teachers and Educational Staff) Room system. The model’s goodness-of-fit shows an NFI (Normed Fit Index) value of 0.632, indicating that the proposed model explains 63.2% of the variance in the data. Novelty: This research presents several significant novelties that contribute to the evaluation of the implementation of the GTK (Teachers and Education Personnel) Room System in primary education. The traditional HOT-Fit (Human, Organization, Technology-Fit) model was enhanced by adding two new variables, Behavioral Intention to Use and Organizational Culture, resulting in a more comprehensive and contextually relevant evaluation framework. The study was conducted within a specific local context, focusing on primary school teachers in Banyumas Regency, Central Java, Indonesia, thereby providing empirical insights into the implementation dynamics at the local level, which have been rarely explored in previous research. The findings reveal that system success is influenced not only by technical factors but also by behavioral dynamics and social contexts, such as organizational culture.
User Experience Improvement (MSMEs and Buyers) Mobile AR Using Design Thinking Methods Dwiyanasari, Desty; Nurhayati, Oky Dwi; Surarso, Bayu; Nugraheni, Dinar
Scientific Journal of Informatics Vol. 12 No. 2: May 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i2.24088

Abstract

Purpose: This research aims to improve the User Experience (UX) of Augmented Reality (AR) mobile applications for MSMEs and buyers through the Design Thinking method. This research solves the problem of suboptimal UX in AR-based mobile applications. This study hypothesizes that the application of Design Thinking can result in significant improvements in the UX of AR mobile applications, which is evidenced by an increase in heuristic evaluation scores. Methods: The Design Thinking approach (Empathize, Define, Ideate, Prototype, Test) is implemented. Data were collected through interviews, observations, and heuristic evaluation questionnaires. Result: Initial heuristic testing showed several usability problems in the developed AR mobile applications, such as Help and Documentation (H10), Recognition Rather than Recall (H6), and Error Prevention (H5). After the application of the Design Thinking method and design iteration, the heuristic testing showed that the results of the evaluation comparison before and after the improvement showed a high effectiveness of the corrective actions taken, with an average decrease in severity score of 37% based on the Nielsen scale (0–4), indicating that the most critical and major issues were successfully reduced to cosmetic or minor levels. Novelty: This research contributes in the form of a practical framework to improve the UX of AR mobile applications for MSMEs and buyers by utilizing the Design Thinking method. The results of this research can be a reference for developers in designing user-friendly AR mobile applications.
Comparison of Ensemble Forest-Based Methods Performance for Imbalanced Data Classification Hasnataeni, Yunia; Saefuddin, Asep; Soleh, Agus Mohamad
Scientific Journal of Informatics Vol. 12 No. 2: May 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i2.24269

Abstract

Purpose: Classification of imbalanced data presents a major challenge in meteorological studies, particularly in rainfall classification where extreme events occur infrequently. This research addresses the issue by evaluating ensemble learning models in handling imbalanced rainfall data in Bogor Regency, aiming to improve classification performance and model reliability for hydrometeorological risk mitigation. Methods: Four ensemble methods: RF, RoF, DRF, and RoDRF were applied to rainfall classification using three resampling techniques: SMOTE, RUS, and SMOTE-RUS-NC. The data underwent preprocessing, stratified splitting, resampling, and 5-fold cross-validation. Performance was evaluated over 100 iterations using accuracy, precision, recall, and F1-score. Result: The combination of DRF with SMOTE-RUS-NC yielded the most balanced results between accuracy (0.989) and computation time (107.28 seconds), while RoDRF with SMOTE achieved the highest overall performance with an accuracy of 0.991 but required a longer computation time (149.30 seconds). Feature importance analysis identified average humidity, maximum temperature, and minimum temperature as the most influential predictors of extreme rainfall. Novelty: This research contributes a comprehensive comparison of ensemble forest-based methods for imbalanced rainfall data, revealing DRF-SMOTE as an optimal trade-off between performance and efficiency. The findings contribute to improved rainfall classification models and offer practical insight for disaster mitigation planning and resource management in tropical regions.
Mental Health Chatbot Application on Artificial Intelligence (AI) for Student Stress Detection Using Mobile-Based Naïve Bayes Algorithm Mariyana, Ekanata Desi Sagita; Novita, Mega; Nur Latifah Dwi Mutiara Sari
Scientific Journal of Informatics Vol. 12 No. 2: May 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i2.24307

Abstract

Purpose: This study aims to design and evaluate a chatbot-based artificial intelligence system to identify stress levels in students using the Naïve Bayes classification method. With increasing mental health concerns among students, early stress detection is considered crucial for timely intervention Methods: This study proposes an AI-based chatbot system to detect student stress levels using a comparative approach between Naïve Bayes and Support Vector Machine (SVM) algorithms. A Kaggle dataset with 15 psychological and academic indicators was preprocessed and balanced using SMOTE. Naïve Bayes showed higher accuracy (90%) than SVM (89%). The trained model was deployed via Flask with Ngrok tunneling and integrated into a Flutter mobile app connected to the Gemini AI API for real-time stress screening. This research offers a practical and scalable solution for early mental health detection in students through intelligent chatbot interaction. Result: The findings show that the Naïve Bayes model achieves a classification accuracy of 90%, slightly surpassing the SVM model, which records an accuracy of 89%. Evaluation through ROC and AUC metrics supports the reliability of Naïve Bayes in detecting stress levels. The integrated chatbot offers a responsive and engaging platform for preliminary mental health assessments. Novelty: This research presents a unique contribution by combining AI-driven stress detection with a real-time chatbot interface, offering an accessible and scalable approach to student mental health support. The integration of machine learning models with conversational AI provides an innovative solution for early intervention. Future developments may involve deep learning and more diverse psychological inputs to further improve accuracy and effectiveness.
Integration of Random Forest, ADASYN, and SHAP for Diabetes Prediction and Interpretation Aulia, Hozana; Wibowo, Adi; Sutrisno, Sutrisno
Scientific Journal of Informatics Vol. 12 No. 2: May 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i2.24314

Abstract

Purpose: Diabetes is a chronic disease with a globally rising prevalence. Early detection of individuals at risk is essential to prevent long-term complications. This study aims to develop a diabetes prediction model that not only achieves high classification accuracy but also provides transparent explanations of the factors influencing its predictions. Methods: The study utilized the Pima Indians Diabetes Dataset, which contains clinical data from 768 female patients aged over 21. The methodology included data preprocessing (handling of missing values and feature engineering, such as the creation of Age_BMI and Glucose_BMI features), a 70:30 train-test split, class imbalance handling using the ADASYN technique, model development using the Random Forest algorithm with hyperparameter tuning via GridSearchCV, and model interpretability analysis using SHAP. Result: The proposed model achieved an accuracy of 79.2% and a recall of 85.2% on the test data. SHAP analysis revealed that Glucose, Age_BMI, BMI, and DiabetesPedigreeFunction were the most influential features in predicting diabetes. Furthermore, the SHAP heatmap indicated that individuals aged 30–50 years with obesity were at the highest risk. These findings align with existing medical literature, reinforcing the role of metabolic and age-related factors in diabetes development. Novelty: This study presents an integrative approach combining class balancing (ADASYN), classification (Random Forest), and model interpretability (SHAP) in a unified framework for diabetes prediction. It emphasizes the importance of transparent model interpretation for healthcare professionals, enabling not only predictive outcomes but also actionable insights into risk factors. The findings support future research opportunities, including the integration of lifestyle variables and external validation using real-world clinical data from diverse populations.
Optimizing LSTM-CNN for Lightweight and Accurate DDoS Detection in SDN Environments Kartadie, Rikie; Kusjani, Adi; Kusnanto, Yudhi; Harnaningrum, Lucia Nugraheni
Scientific Journal of Informatics Vol. 12 No. 2: May 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i2.24531

Abstract

Purpose: This study optimizes the LSTM-CNN model to detect Distributed Denial of Service (DDoS) attacks in Software-Defined Networking (SDN)-based networks and improves accuracy, computational efficiency, and class imbalance handling. Methods: We developed an Improved LSTM-CNN by removing the Conv1D layer, reducing LSTM units to 64, and using 21 features with a 5-timestep approach. The InSDN dataset (50,000 samples) was preprocessed with one-hot encoding, MinMaxScaler normalization, and sequence formation. Class imbalance was managed using class weights (0:2.0, 1:0.5) instead of SMOTE, with performance compared against Baseline LSTM-CNN and Dense-only models optimized with the Sine Cosine Algorithm (SCA). Result: The Improved LSTM-CNN achieved 0.99 accuracy, 0.93 F1-score for Benign traffic, and 1.00 for Malicious traffic, with ~25,000 parameters and 125 ms inference time on Google Colab. It outperformed Baseline LSTM-CNN (0.08 accuracy) and was more efficient than Dense-only (46,000 parameters), with a false positive rate of ~1%. Novelty: This research presents a lightweight, efficient DDoS detection solution for SDN, leveraging temporal modeling and class weights, suitable for resource-constrained controllers like OpenDaylight or ONOS. However, its generalization is limited by dataset diversity, necessitating broader validation.
Evaluation of Ridge Classifier and Logistic Regression for Customer Churn Prediction on Imbalanced Telecommunication Data Rofik, Rofik; Unjung, Jumanto
Scientific Journal of Informatics Vol. 12 No. 2: May 2025
Publisher : Universitas Negeri Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15294/sji.v12i2.24620

Abstract

Purpose: Customer churn is a crucial issue for companies, especially those in the telecommunications sector, as it has a direct impact on revenue and new customer acquisition costs. The purpose of this research is to create a customer churn prediction model through performance comparison between the Logistic Regression algorithm and Ridge Classifier, considering the effect of data balancing. Methods: This study developed a churn classification model by comparing the Logistic Regression and Ridge Classifier algorithms in three scenarios: without data balancing, balancing using SMOTE, and balancing using GAN. The dataset used was Telco Customer Churn from Kaggle. Model evaluation was performed using a confusion matrix with accuracy, precision, recall, and F1-score metrics, with a primary focus on the accuracy metric. Result: The results show that data balancing using SMOTE and GAN does not improve model accuracy. The highest accuracy was achieved by the Ridge Classifier without data balancing, at 82.47%, followed by Logistic Regression at 82.25%. However, the recall and F1-score metrics improved when using SMOTE. The highest recall was achieved by Ridge Classifier at 75.34% and Logistic Regression at 75.07% in the SMOTE 50:50 scenario. The highest F1-score was also achieved by Ridge Classifier at 64.76% and Logistic Regression at 64.68% followed by the SMOTE 50:30 scenario. Meanwhile, the precision metric tends to decrease after data balancing. Novelty: The uniqueness of this study lies in the comparison of the performance of the Ridge Classifier and Logistic Regression in data balancing scenarios using SMOTE and GAN, which has not been widely discussed in previous studies. The main findings show that the highest accuracy is achieved when the Ridge Classifier model uses original data or without applying SMOTE or GAN data balancing. However, data balancing using SMOTE has been proven to significantly improve the recall and F1-score metrics.