INTENSIF: Jurnal Ilmiah Penelitian dan Penerapan Teknologi Sistem Informasi
Not a member yet
174 research outputs found
Sort by
Enhancing SVM-Based Classification Performance on Indonesian Sentences through TF-IDF and Directional Augmentation
Full text linkBackground: The distinction between standard and non-standard Indonesian sentences is traditionally well-defined, yet the ubiquity of digital communication has increasingly blurred these boundaries. This convergence introduces significant lexical ambiguity in formal contexts, complicating the performance of automated text classification systems. Objective: This study aims to enhance the robustness of Support Vector Machine (SVM) classification by addressing these linguistic irregularities through TF-IDF vectorization and a targeted directional augmentation strategy. Methods: A corpus comprising 5,394 labeled sentences was processed under a strict anti-leak grouping strategy to rigorously prevent semantic leakage between training, validation, and testing sets. To resolve decision boundary overlaps often missed by the baseline model, manual directional augmentation was applied, specifically targeting ambiguous sentence structures to enrich the training distribution and linguistic diversity. Results: The experiments demonstrated that directional augmentation significantly refined the model\u27s decision margins. While the baseline model achieved a test accuracy of 94.39%, the augmented approach substantially improved generalization capabilities across unseen groups, elevating validation accuracy from 96.11% to 97.39% and test accuracy to 96.16%. Conclusion: These findings substantiate that structurally enriching the dataset effectively mitigates overfitting and improves sensitivity. However, given the scalability constraints of manual intervention, future research should prioritize automated augmentation techniques and contextual embeddings to handle deep linguistic nuances further
Evaluation of Mobile Application Service on User Loyalty Using Expectation Confirmation Model
Full text linkBackground: Mobile-based Academic Information Systems (AIS) have become essential for improving accessibility and efficiency in higher education. UIN Jakarta’s AIS Mobile aims to support academic activities; however, user loyalty remains low, as many students prefer accessing services via the web platform. Objective: This study evaluates AIS Mobile services and identifies key factors influencing user loyalty using an extended Expectation Confirmation Model (ECM). Methods: A quantitative approach was employed, involving 334 respondents selected through purposive sampling. Data were collected via an online questionnaire and analysed using Partial Least Squares Structural Equation Modelling (PLS-SEM) with SmartPLS 4.0.9.3. The proposed model integrates ECM constructs—confirmation, perceived usefulness, satisfaction, and continuance intention—with additional variables: system quality, information quality, trust, habit, and loyalty. Results: Findings indicate that eight hypotheses were supported, confirming significant relationships among confirmation, perceived usefulness, trust, habit, and continuance intention in shaping loyalty. Satisfaction, however, showed no significant effect on continuance intention. The model demonstrates strong explanatory power, with R² values of 0.737 for continuance intention and 0.726 for satisfaction. Habit exhibited the largest effect size, emphasizing its role in sustaining usage. Implications: To enhance user loyalty, developers should prioritize improving system reliability, security, and usability while fostering habitual engagement through intuitive design and personalized features. These insights provide actionable strategies for strengthening AIS Mobile adoption in Islamic higher education contexts
Word Stemming of Lampung Dialect Nyo using N-Gram Stemming
Full text linkBackground: Previous translation systems for the Lampung dialect of nyo to Indonesian achieved bilingual evaluation understudy (BLEU) scores below 40%, primarily due to challenges in processing affixed words. Objective: This research aims to perform stemming on affixed words in the Lampung dialect of nyo to enhance the performance of the translation system. Methods: We developed an n-gram stemming approach that reduces affixed words to their base forms by measuring similarity between n-grams using the Dice coefficient method. When similarity exceeds a specified threshold, the system identifies the corresponding base word. Results: Using a dataset of 700 words from the Lampung dialect of nyo, we constructed a comprehensive stemmer covering all affix variations. The optimal threshold was determined to be 0.5, achieving bigram accuracy of 93.86% and trigram accuracy of 89.14%. These accuracy levels demonstrate the method\u27s effectiveness in identifying base word forms, which directly impacts translation quality improvement. Conclusion: N-gram stemming with a 0.5 threshold effectively processes the Lampung dialect of nyo morphology and shows potential for enhancing translation accuracy. This work represents the first comprehensive stemming system specifically designed for the Lampung dialect of nyo, contributing to the development of natural language processing tools for underrepresented regional languages in Indonesia.
Enhancing Tourist Experiences in North Toraja through K-Means Clustering-Based Recommendation System
Full text linkBackground: North Toraja in South Sulawesi, Indonesia, is a culturally rich region with high tourism potential due to its unique traditions. The government has invested in infrastructure to boost tourism and regional income (PAD), but has insufficiently used information systems for promotion. An innovative system that can assist tourists in navigating the diverse attractions in North Toraja based on their interests needs to be developed. Objective: This research aims to develop a recommendation system for tourist attractions in North Toraja using K-means Clustering and the Similar Characteristics Method. Methods: We used Orange Data Mining to perform K-means clustering, and then used similarity-based methods to determine the closeness of characteristics among attractions. The system analyzes based on the fields of cultural, geographical, facility, and landscape features, resulting in four distinct clusters. The clusters were defined as three tourist attractions in cluster C1, eleven in C2, four in C3, and fourteen in C4. We also developed a system interface that allows travelers to input preferences, view personalized recommendations, and access detailed information. The system\u27s novelty lies in its specific application of K-Means Clustering to leverage these local attributes for granular categorization for effective promotion of North Toraja\u27s diversity. Conclusion: Our approach effectively groups attractions with similar characteristics, enhancing exploration based on user interests. The high altitude and similar geographical features of North Toraja result in attractions that share natural characteristics, making this system an advancement in technology-driven tourism solutions.
Security Assessment Based on OWASP Top 10 Using SonarQube and ZAP on Export and Import Applications in the LNSW
Full text linkBackground: The advancement of information and electronic systems has significantly transformed export and import processes. In Indonesia, the Lembaga National Single Window (LNSW) plays a pivotal role in facilitating international trade by integrating procedures and information related to exports, imports, and document flows. Objective: This study aims to assess the security of LNSW’s export and import application by identifying vulnerabilities based on the Open Web Application Security Project (OWASP) Top 10 framework. It also compares the effectiveness of Static Application Security Testing (SAST) using SonarQube and Dynamic Application Security Testing (DAST) using ZAP (Zed Attack Proxy) in detecting various types of vulnerabilities. Methods: The analysis involved the use of SonarQube for source code scanning and ZAP for runtime testing. Each detected vulnerability was evaluated using the Common Vulnerability Scoring System (CVSS) to determine its severity level. Recommended mitigation strategies were provided accordingly. Results: A total of eight vulnerabilities were identified, comprising two High-severity and six Medium-severity issues. SonarQube proved more effective in detecting Identification and Authentication Failures (three instances), while ZAP excelled in identifying Vulnerable and Outdated Components (two instances). Notably, each tool uncovered four unique types of vulnerabilities that the other did not detect. Conclusion: These findings highlight the practical benefits of combining SAST and DAST techniques. By integrating both approaches, organizations can achieve a more comprehensive and reliable security assessment, ultimately leading to more resilient software systems.
Analysis of Information Systems Acceptance and Success Models in Higher Education
Full text linkBackground: The integration of Information Systems (IS) in higher education has transformed interactions among students, lecturers, and administrative staff, making system acceptance and success essential for effective academic processes. Various evaluation frameworks have been developed, with the DeLone and McLean Information System Success Model being one of the most widely applied. Objective: This study aims to analyze factors influencing the adoption of academic information systems in higher education using the DeLone and McLean model and to evaluate system success from the perspectives of lecturers, students, and administrative personnel. Methods: A quantitative research approach was employed using questionnaire-based data collection. Data analysis was conducted using SmartPLS 3.0 to assess validity, reliability, and structural relationships among variables. A total of 252 respondents were selected using the Slovin formula and proportional stratified random sampling. The evaluated constructs included system quality, information quality, service quality, system use, user satisfaction, and benefits. Results: The results show that system quality, information quality, and service quality have a positive and significant effect on system use and user satisfaction. Furthermore, system use and user satisfaction contribute to perceived net benefits, such as improved learning outcomes, increased management efficiency, and academic productivity. High service quality also supports continued system usage. All measurement constructs met validity and reliability criteria, with loading factors above 0.7 and Average Variance Extracted (AVE) values exceeding 0.50. Conclusion: In conclusion, the DeLone and McLean model effectively explains academic information system success in higher education, highlighting the importance of system quality, user satisfaction, and generated benefits
Uncovering Key Topics in Indonesian Political Discourse Through Twitter Analysis After the 2024 Presidential Inauguration Using Clustering methods
Full text linkBackground: Social media, especially Twitter, plays a key role in political discourse, shaping public opinion. In Indonesia, the 2024 presidential Inauguration , with candidates Prabowo Subianto and Gibran Rakabuming Raka, has generated significant online conversations. Understanding public sentiment and identifying key topics is urgent for analyzing and grouping these discussions, offering insights into political views. Objective: The purpose of this research is to analyze Twitter conversations surrounding the 2024 Indonesian presidential election. The goal is to identify the main topics in these conversations and assess the effectiveness of different clustering algorithms in grouping similar tweets. Methods: This study applies a quantitative approach, using a dataset of 29,905 tweets collected from October 20 to October 25, 2024. The method includes text preprocessing, such as tokenization, stemming, and word weighting. PCA is used for dimensionality reduction. The clustering algorithms K-means, DBSCAN, PAM, and Agglomerative Hierarchical are employed, with performance evaluated based on the Silhouette Score. Results: The results reveal that the Agglomerative Hierarchical Clustering algorithm with Ward linkage and two PCA components produced the highest Silhouette Score of 0.8018. The clustering identified three distinct topics: political leadership, work and collaboration, and unity. Conclusion: This research successfully identified key discussion topics in Twitter conversations about the 2024 Indonesian presidential election. The Agglomerative Hierarchical method with Ward linkage was the most effective clustering algorithm. These findings offer valuable insights into public opinion, and future studies could expand to other social media platforms or investigate the relationship between sentiment and political outcomes
Evaluating YOLOv8-Based Distance Estimation: A Comparison of OpenCV and Coordinate Attention Weighting in Blind Navigation Systems
Full text linkBackground: Recent developments in assistive technologies for the visually impaired have increasingly utilized computer vision techniques for real-time distance estimation. However, challenges remain in balancing accuracy, latency, and robustness under dynamic environmental conditions. Objective: This study aimed to evaluate and compare the performance of OpenCV and Coordinate Attention Weighting (CAW) models for distance estimation in blind navigation systems, particularly focusing on their effectiveness in real-time scenarios. Methods: A quantitative experimental study was conducted using an image dataset labeled with actual distances. The baseline performances of OpenCV and CAW were measured and compared. Subsequently, targeted optimizations were applied to the OpenCV model, including adaptive image filtering, hyperparameter tuning, and integration of a Kalman filter. Results: Initial evaluation showed that CAW achieved a higher baseline accuracy of 88% compared to OpenCV. However, after optimizations, OpenCV’s accuracy improved by 15%, reaching approximately 85%. Additionally, the optimized OpenCV model demonstrated reduced latency, outperforming CAW in real-time detection speed. Under varying lighting and motion conditions, OpenCV also exhibited superior robustness compared to CAW. Conclusion: The findings suggest that with proper optimization, OpenCV can match or exceed CAW in key performance aspects, making it a viable and efficient alternative for real-time distance estimation in blind navigation systems. Future research should explore further model integration and hardware acceleration for deployment in wearable devices
Extending the Expectation Confirmation Model to Examine Continuous Use Mobile Banking: Security, Trust, and Convenience
Full text linkBackground: Mobile banking adoption continues to grow, but user retention remains a challenge. Understanding the factors influencing continuance intention is crucial for improving long-term engagement. Prior research highlights the importance of confirmation, perceived usefulness, security, satisfaction, trust, and convenience, yet their interrelationships require further exploration. Objective: This study examines key determinants of users\u27 intention to continue using mobile banking services, focusing on how confirmation, perceived usefulness, security, satisfaction, trust, and convenience influence this decision. Methods: A quantitative study was conducted using structural equation modeling (SEM) to analyze relationships among these factors. Data were collected from mobile banking users and assessed for statistical significance. Results: Confirmation significantly impacts perceived usefulness (0.576) and satisfaction (0.527). Perceived usefulness influences satisfaction (0.289) and continuance intention (0.396), while satisfaction also affects continuance intention (0.240). Trust plays a role (0.211), and perceived security strongly influences trust (0.651). Perceived convenience also impacts continuance intention (0.304), emphasizing its importance in user experience. Conclusion: Confirmation and security are critical for satisfaction and trust, which drive continued mobile banking use. Strengthening security, improving perceived usefulness, and fostering trust can enhance user retention. Future studies should explore additional variables, test the model across demographics, and assess the impact of emerging technologies like AI and blockchain. Longitudinal and experimental research may offer deeper insights into these evolving relationships
Optimizing Rice Plant Disease Classification Using Data Augmentation with GANs on Convolutional Neural Networks
Full text linkBackground: Rice disease classification using CNN models faces challenges due to limited data, particularly in minority classes, and inconsistent image quality, which affect model performance. Data augmentation techniques can potentially enhance classification accuracy by improving data diversity and quality. Objective: This study evaluates the effectiveness of data augmentation techniques, specifically classical augmentation and Deep Convolutional Generative Adversarial Networks (DCGAN), in improving CNN performance for rice disease classification. Methods: A quantitative study was conducted using four CNN training scenarios: no augmentation, classical augmentation, DCGAN augmentation, and a combination of both. Model accuracy was analyzed to determine the impact of each augmentation technique. Results: The baseline CNN model achieved an accuracy of 91.88%. Classical augmentation improved accuracy by 2.56%, while DCGAN augmentation led to a 5.44% increase. The combination of classical augmentation and DCGAN yielded the highest accuracy of 98.13%. Conclusion: Data augmentation significantly enhances CNN performance in rice disease classification, with the combined approach of classical augmentation and DCGAN proving to be the most effective. These findings highlight the importance of augmentation techniques in addressing data limitations and improving classification accuracy. Future research should explore additional augmentation strategies and test the model across different datasets to further validate its effectiveness