Jurnal Sistemasi (OJS FTIK - UNISI, Fakultas Teknik dan Ilmu Komputer Universitas Islam Indragiri)
Not a member yet
946 research outputs found
Sort by
Implementation of the K-Means Algorithm for Clustering Zoning of Natural Disaster-Prone Areas in Pati Regency
Full text linkBased on data from the Central Bureau of Statistics (BPS) of Pati Regency, during the 2018–2022 period the region frequently experienced natural disasters, particularly floods and landslides, across many areas.The high frequency of these disasters and the lack of proper mapping of affected areas present challenges that need to be addressed effectively. By utilizing technology and the K-Means clustering method, this study proposes an alternative solution to identify and map areas that are vulnerable to natural disasters. The results of the analysis indicate that Pati Regency can be divided into three clusters: Cluster 1 represents highly disaster-prone areas, accounting for 42.86% (9 regions); Cluster 2 represents disaster-prone areas, accounting for 19.04% (4 regions); and Cluster 3 represents areas with low disaster vulnerability, accounting for 38.1% (8 regions). The visualization results are presented as a classification map of regions based on their disaster vulnerability levels. This map can serve as a reference for local governments and relevant institutions in formulating more targeted and effective disaster mitigation policies
Web-based Information System for Research and Community Service (PkM) Submissions at LPPM, Nahdlatul Ulama University Lampung
Full text linkManagement of research and community service (PkM) activities in higher education institutions requires an integrated information system to enhance the effectiveness, transparency, and accountability of processes. The LPPM at Nahdlatul Ulama University Lampung still faces challenges in managing proposal submissions, reviewer assignments, the review process, and the handling of research and PkM reports, which are not yet fully integrated into a single system. This study aims to develop and implement a web-based Research and PkM Submission System for LPPM at Nahdlatul Ulama University Lampung. The research adopts the Waterfall model, covering the stages of analysis, design, implementation, testing, and maintenance. The system is built using PHP and MySQL, supported by HTML, CSS, and Bootstrap technologies. To ensure functionality meets the specified requirements, verification was conducted using Blackbox Testing. In addition to functional testing, application performance was evaluated by measuring system response times for login, proposal storage, document uploads, and dashboard display on both desktop and mobile devices. Test results indicate that the platform responds within an average of 1–3 seconds for key functions and demonstrates stable performance without technical issues throughout the testing period. The findings show that the application successfully integrates the entire workflow of proposal submission, reviewer assignment, proposal and report review processes, and monitoring of research and PkM status into a single platform. Testing results confirm that all system modules function correctly and align with the testing scenarios. Therefore, this system is considered suitable for implementation as a supporting management tool for research and PkM activities at LPPM Nahdlatul Ulama University Lampun
Optimization of the Linear Regression Algorithm using GridSearchCV for Rice Crop Production Prediction
Full text linkRice production in Central Java Province fluctuates annually, affecting food security and agricultural output distribution. Therefore, accurate prediction methods are essential to assist stakeholders in agricultural planning and strategic decision-making. This study applies the Linear Regression algorithm to predict rice production based on historical data from 2014 to 2023 obtained from the official website of the Central Java Provincial Agriculture and Plantation Office. The model is developed using multiple linear regression with variables including planted area, harvested area, and productivity. The novelty of this study lies in the structured application of hyperparameter tuning using GridSearchCV to optimize Linear Regression performance, as well as the integration of a preprocessing pipeline based on data distribution stabilization to improve accuracy and model generalization. The research process includes data collection, preprocessing, modeling, optimization, model evaluation, and deployment as a web-based application using Streamlit Cloud. GridSearchCV optimization results indicate a cross-validation accuracy of 98.26%, confirming the model’s strong predictive capability. Model evaluation shows an R² value of 0.9754, with MAE of 0.0957, MSE of 0.0307, and RMSE of 0.1753, indicating low prediction errors and stable model performance. The optimized model is implemented as a web application via Streamlit Cloud, enabling direct use by end-users. For future research, it is recommended to incorporate additional variables such as rainfall, temperature, and rice variety, or to compare performance with other algorithms such as Random Forest, Support Vector Regression, or Long Short-Term Memory (LSTM) to further enhance prediction accuracy
Development of an Intelligent Platform for Drug and Food Interaction Analysis using a Combination of Fuzzy Logic and Certainty Factor
Full text linkInteractions between drugs and food are a critical public health issue, as they can cause unwanted side effects or reduce the effectiveness of treatments. Unfortunately, awareness of these potential interactions among the Indonesian population remains low, while existing platforms generally focus only on drug–drug interactions. This study aims to develop an intelligent platform for analyzing drug–food interactions by combining fuzzy logic and certainty factor (CF) methods. Fuzzy logic is employed to handle uncertainty in interaction data, while the certainty factor enhances confidence levels based on clinical literature and expert knowledge. Drug–food interaction data were collected from validated sources and modeled using fuzzy membership functions, IF–THEN rule-based reasoning, defuzzification processes, and integration with CF. The web-based system was evaluated through accuracy testing and usability assessment using the System Usability Scale (SUS). Accuracy tests conducted on 50 interaction scenarios demonstrated a 100% match with clinical references, while the SUS evaluation involving 100 respondents yielded an average score of 77.44, falling into the “Acceptable” category and approaching “Good Usability.” These results indicate that the platform has the potential to serve both as an educational tool and as a practical aid for the public to enhance self-management of health, while also supporting government health programs
Expert System for Mental Health Disordes in Women and Children based on Android using the Certainty Factor Method
Full text linkThis study focuses on the limited access to mental health services in Gorontalo and the social stigma surrounding mental health disorders, which discourages women and children from seeking help. The main problem addressed in this research is the management of mental health disorders, particularly the complexity of the initial diagnostic process. The study aims to develop an expert system using the Certainty Factor method to assist in the early diagnosis of mental health disorders in women and children, specifically Depression, Anxiety, and Stress disorders. The research employs a Research and Development (R&D) approach with qualitative methods, including interviews with experts such as psychologists to obtain a knowledge base comprising symptoms, Measure of Belief, and Measure of Disbelief values. The expert system is implemented on the Android platform, facilitating user access to early diagnostic services. Results from the Certainty Factor calculations on user data indicate that the early diagnosis confidence levels are 97.04% for Depression, 63.11% for Anxiety, and 59.72% for Stress. The highest value is observed for Depression, suggesting that the symptoms selected by users most strongly indicate this disorder, with the highest confidence level across all Depression symptoms. Both manual calculations by experts within the system and Black Box testing confirm that the Certainty Factor method can effectively support early diagnosis of mental health disorders. The study concludes that the expert system using the Certainty Factor method is effective and can be implemented as an early mental health detection tool. The strength of this research lies in the integration of qualitative expert knowledge with mobile technology implementation, providing a practical and easily accessible solution for the community
The Influence of Experience-Centric IT Governance on Digital Ethics Perception in Social Commerce
Full text linkSocial commerce is a rapidly growing new form of e-commerce that integrates social functions into digital buying and selling activities. However, studies explicitly examining the influence of Experience-Centric IT Governance (ECITG) approach on digital ethics perceptions are limited. This study aims to evaluate ECITG's influence on users perceptions of digital ethics in social commerce. A quantitative approach was used to measure ECITG through five variables: Governance Responsiveness, Transparency and Trust, Experience Personalization, User Participation, and User Empathy. Digital ethics perceptions encompass five variables: Platform Accountability, System Fairness, Data Privacy, Consumer Protection, and Algorithm Transparency. Data were gathered from 100 respondents and processed using partial least squares–structural equation modeling (PLS-SEM) through measurement and structural. The results of the analysis indicate that several ECITG dimensions significantly shape digital ethics perception, particularly governance responsiveness, transparency, and user participation, while personalization and empathy demonstrate weaker influence. Overall, these results confirm that the ECITG approach has a statistically significant impact on digital ethics perceptions in the context of social commerce, though its strength varies across dimensions. The study contributes to the development of experience-based IT governance models and offers practical insights for platform managers to improve ethical transparency and user trust
Development of Hybrid K-Means DBSCAN Algorithm for Optimization of Landslide-Prone Area Clusters based on Web-GIS
Full text linkLandslides represent one of the major geological hazards in West Java Province, posing serious impacts on social life, economic activities, and public infrastructure. A key challenge in landslide mitigation lies in the inaccuracy of spatial and temporal classification of landslide-prone areas, as well as the limitations of single-method approaches in disaster data analysis. This study aims to develop a data-driven classification model for landslide-prone areas using a hybrid clustering approach that combines the K-Means and DBSCAN algorithms. The dataset consists of landslide incident records from 2020 to 2024 and administrative spatial data at the regency/city level. The analysis stages include data integration and normalization, statistical exploration, the application of K-Means clustering as a global segmentation framework, and DBSCAN for identifying local patterns and outliers. Model validation was conducted using internal evaluation metrics, yielding a Silhouette Coefficient of 0.448 and a Davies–Bouldin Index of 0.602, indicating that the hybrid method provides superior performance in terms of cluster compactness and separation. The classification results are visualized through an interactive Web-GIS platform developed using Streamlit and Folium, enabling users to select specific years and classification methods while displaying mitigation strategies based on risk categories. This study concludes that the hybrid clustering approach enhances the accuracy of landslide-prone area classification and makes a significant contribution to the provision of more adaptive and practical spatial information to support mitigation policy decision-making in landslide-vulnerable regions
Rice Plant Disease Detection System based on Leaf Image using Web-based CNN Algorithm
Full text linkRice (Oryza sativa) plays a crucial role as a major staple food commodity. However, diseases such as Bacterial Blight, Brown Spot, and Leaf Blast can cause significant crop losses. Current manual identification methods have limitations due to high subjectivity and long diagnosis time. This study proposes a web-based automatic detection system using a Convolutional Neural Network (CNN). The dataset was obtained from Kaggle and consisted of 2,800 images evenly distributed across four classes (700 images per class). The data were split using an 80:20 ratio for training and validation sets, followed by preprocessing steps including resizing to 224×224 pixels and data augmentation. The CNN architecture was designed with four convolutional blocks and optimized using the Adam optimizer. Training for 50 epochs achieved an accuracy of 77.50%, precision of 82.98%, recall of 77.50%, and an F1-score of 72.84%. Based on the confusion matrix analysis, the model performed very well in detecting Bacterial Blight and Brown Spot but still faced difficulties in identifying the Leaf Blast class. Overall, the developed system has the potential to serve as a decision-support tool for farmers, although further performance improvements are required, particularly for detecting specific disease variants
An Automated System for Detecting and Improving Academic Text Politeness Using IndoBERT and IndoT5
Full text linkThe increasing use of digital communication in academic interactions between students and lecturers is often not accompanied by consistent application of language politeness norms, potentially affecting the effectiveness of academic interactions. To date, efforts to enhance language politeness have predominantly relied on manual and subjective evaluation. This study aims to develop an automated system for detecting and improving politeness in Indonesian academic text communication. The proposed approach integrates IndoBERT as a classification model to identify levels of text politeness and IndoT5 as a generative model to transform sentences identified as impolite into more appropriate academic forms. The dataset consists of 6,230 labeled sentences collected through Google Forms, TikTok, and additional synthetic data generated using ChatGPT. Experimental results show that the IndoBERT model achieves an accuracy of 97.11% in classifying academic text politeness, while IndoT5 is capable of transforming impolite sentences into more appropriate academic expressions, as demonstrated by evaluations using BLEU, ROUGE, and METEOR metrics. This study results in an integrated deep learning–based system capable of automatically detecting and improving academic text politeness within a unified processing framework
Detecting Chili Ripeness Using YOLOv11
Full text linkThis study aims to develop a deep learning-based chili ripeness detection system using the YOLOv11 model. Chili ripeness is classified into three categories: unripe, semi-ripe, and ripe. The dataset consists of 150 original images, which were expanded to 300 images to increase data variation. Model training was conducted using the Roboflow platform, while accuracy testing was performed in Google Colab through an image upload-based processing method. The experimental results show that the model achieved an accuracy of 93.94%, with a precision of 94.21%, recall of 93.94%, and an F1-score of 93.94% on the test dataset. This system is expected to support the automation of chili sorting based on ripeness levels