Publication Search

The coffee shop industry in Indonesia is experiencing rapid growth that requires business owners to optimize data-driven strategies. This study aims to analyze customer preferences at Semanis Coffee and Resto using data mining methods  to support more effective business decision-making. The method used is Market Basket Analysis with the FP-Growth algorithm for association rule mining and the K-Means algorithm for customer segmentation. The research data consists of 672 sales transactions during the March-May 2025 period. The results of the association analysis with a minimum support of 0.004 and a minimum confidence of 0.2 resulted in five valid rules with a lift ratio above 1. The strongest rule is the combination of Americano→Milk Choco with a confidence of 42.9% and an elevator ratio of 5.229, indicating a strong linkage between products. The most popular products are Milk Choco (10.8%) and Americano (8.5%). Customer segmentation analysis identified three clusters: Cluster 0 (Loyal Customers) 80% with high frequency but low transaction value; Cluster 1 (Occasional Customers) 10% with low activity; and Cluster 2 (Large Buyers) 10% with high transaction value but low frequency. This study concludes that product bundling strategies, loyalty programs, reactivation campaigns, and premium services can be applied to increase the effectiveness of coffee shop businesses.

Dengue fever is a disease commonly found in tropical and subtropical regions. This disease can cause severe symptoms, such as very high fever, accompanied by nausea, vomiting, headache, abdominal pain, and leukopenia (decrease in white blood cells). This infectious disease, known as dengue hemorrhagic fever (DHF), is a viral infection transmitted by the Aedes Aegyppti mosquito. This study aims to classify dengue-prone areas using the K-Means Algorithm, and to classify the factors that cause dengue in Purwodadi District, Grobogan Regency. The clustering results using the K-Means algorithm with Rapidminer tool from 266 data produced 3 clusters: cluster 0 (blue) with 138 patients dominated by Kuripan, Purwodadi, Ngambak villages, cluster 1 (green) with 31 patients in Ngraji, Nambuhan, Cingkrong villages, and cluster 2 (orange) with 97 patients in Danyang, Kalongan, Pulorejo villages. This study is expected to provide additional information for stakeholders in controlling dengue cases and increase awareness of the importance of environmental cleanliness as a preventive measure.

This study aims to identify domestic visitor segments in Bromo Tengger Semeru National Park (TNBTS), Indonesia, based on travel characteristics and consumption patterns to support the development of quality tourism in protected areas. Using snowball sampling, 283 domestic visitors was analysed by Two-Step Cluster Analysis in SPSS by integrating length of stay, activity preferences, and expenditure patterns. The results reveal a two-cluster solution as the most optimal segmentation, supported by the highest ratio of distance measures, with cluster quality rated as fair (silhouette = 0.20). Cluster 1 (39.2%) represents short-stay, lower-spending visitors who primarily seek iconic experiences (“Sunrise Seekers”), while Cluster 2 (60.8%) reflects longer-stay, higher-spending visitors who prefer village tourism activities (“Village Experience Seekers”). The strongest differentiating variables across segments are length of stay, activity preference, expenditure style, and age, whereas gender, education level, origin, and travel companions have limited role in segment separation. This study contributes empirical evidence of data-driven visitor segmentation in a conservation-based ecotourism destination within a volcanic national park, extending prior expenditure-focused profiling by integrating length of stay and activity preferences to capture visitor heterogeneity more comprehensively.

Transaction logs in online retail provide opportunities for data-driven customer segmentation. This study segments customers at two scopes global (all countries) and United Kingdom (UK) using Recency, Frequency, and Monetary (RFM) features derived from the Online Retail transaction dataset. After cleaning cancellations and invalid records, RFM variables are computed per customer and normalized. K-Means clustering is applied separately for global and UK data, while the number of clusters is selected via the elbow criterion and validated using internal indices. The best configuration for both scopes yields five clusters, with moderate separation quality based on the silhouette score. Cluster profiling indicates distinct groups ranging from low-frequency low-spending customers to highly frequent high-spending customers. The comparison between global and UK segmentation shows similar structural patterns, yet different proportions across segments, supporting targeted retention and value-driven marketing actions.

This study presents ReflectAI, a web-based system designed to automate the creation of teaching materials tailored to students' learning styles using behavior data from a Learning Management System (LMS). Student digital activity data—such as logins, material access, forum participation, assignment submission, and quiz results—are extracted and processed using a Hierarchical Clustering algorithm to categorize students into three learning styles: visual, auditory, and kinesthetic. Based on the clustering results, the system automatically generates personalized learning modules using generative AI (ChatGPT API), aligned with each student's learning preferences. Employing a data-driven system development approach, the system was tested with data from 230 students in a mathematics course. The results show diverse learning style distributions and relevant, tailored content generation. ReflectAI is designed to reduce teachers’ administrative workload and enhance personalized and adaptive learning. This system contributes to educational transformation through deep, data-driven technology integration.

The increasing use of Learning Management Systems (LMS) in higher education generates large amounts of student activity data that have the potential to provide deeper insights into learning processes. However, in practice, these data are still rarely analyzed systematically to understand variations in students’ learning activity patterns, limiting their practical use in supporting teaching and learning. This study aims to explore students’ learning activity patterns in an LMS using a clustering approach based on activity data.This research utilizes the publicly available Open University Learning Analytics Dataset (OULAD), focusing on a single course and a single academic term. LMS activity data were processed through data cleaning and feature extraction, followed by student clustering using the K-Means algorithm. The quality of the clustering results was evaluated using the Silhouette Score, and visual analysis was applied to support the interpretation of the results.The results indicate that students’ learning activities can be grouped into two main patterns, namely a group of students with high learning activity and a group with lower or moderate activity levels. These findings highlight the existence of heterogeneous learning behaviors among students, even within the same learning context.The identified learning activity patterns provide an initial foundation for utilizing LMS data to monitor student engagement and to support the development of more responsive, data-driven learning approaches in higher education.

Savings and Loan Cooperatives (KSP) play a vital role in expanding community access to capital, especially within the informal sector. Nevertheless, non-performing loans remain a persistent challenge that can threaten liquidity and long-term institutional sustainability. KSP CU Mera Ndi Ate faces similar issues, which are assumed to stem not only from administrative weaknesses but also from members’ perceptions and behavioral factors. This research aims to examine the potential causes of non-performing loans through text-based sentiment analysis using an unsupervised learning approach. A quantitative method with a data mining framework was applied. Data were gathered through interviews, observations, documentation, and 200 customer opinion texts processed using the Orange Data Mining application. The analytical stages included preprocessing, corpus development, feature extraction, sentiment clustering, and visualization. Because the dataset lacked predefined labels, unsupervised learning was used to identify naturally emerging sentiment patterns. Findings reveal a predominance of critical sentiments related to credit assessment procedures and service quality. The highest sentiment score (75) concerned insufficient creditworthiness evaluation, followed by concerns about service efficiency (66.6667). These insights suggest that improving assessment accuracy and service quality may help reduce non-performing loans.

Graph theory as a branch of discrete mathematics has experienced significant development in its application to modern complex network systems, particularly in digital social networks and transportation systems. This research aims to analyze fundamental concepts of graph theory, examine characteristics of cycle detection algorithms along with their computational complexity, investigate their application in digital social network analysis, and explore their implementation in digital transportation system optimization. The research method employs a qualitative approach with library research focusing on scientific literature from 2020-2025 period from accredited academic databases such as Scopus, Web of Science, and IEEE Xplore, utilizing thematic analysis techniques to identify meaningful patterns from the examined literature. Research findings indicate that fundamental graph theory concepts including vertices, edges, and graph classifications form the foundation for relational structure modeling. Cycle detection algorithms such as Depth-First Search, Union-Find, and Tarjan demonstrate effectiveness with O(V+E) complexity for large-scale graphs. Applications in digital social networks facilitate community identification through Multi-View Clustering, centrality analysis for influencer detection, and understanding viral information dissemination patterns. Implementation in digital transportation systems demonstrates route planning optimization using Dijkstra and Bellman-Ford algorithms, vulnerability analysis through articulation point and bridge identification, and bottleneck detection with betweenness centrality. The research concludes that integration of graph theory in discrete mathematics education enhances critical thinking skills and real-world application understanding, with recommendations for algorithm development for massive dynamic graphs and machine learning integration in graph algorithm optimization.

This study explores the role of visual analytics in enhancing decision-making processes within creative industries, focusing on its application to large-scale multimedia datasets. Visual analytics integrates interactive visualization techniques with computational algorithms, enabling users to explore complex datasets intuitively and derive actionable insights. The research centers on the design and implementation of interactive dashboards tailored to the creative sector, particularly film, music, and advertising industries, to facilitate real-time data exploration. The study also investigates the usability of these tools through expert-based evaluations, aiming to assess their effectiveness in supporting informed and timely decision-making. The findings reveal that interactive visualizations significantly improve insight discovery and pattern recognition, enabling decision-makers to uncover hidden trends in large multimedia datasets. However, challenges related to scalability, user acceptance, and real-time processing were encountered during the implementation phase. The research highlights the practical benefits of integrating visual analytics into industry workflows, which include enhanced content creation, audience engagement, and strategic planning. Furthermore, the study identifies key visual analytics techniques such as dynamic dashboards, pattern recognition, data mining, and clustering, which are essential for analyzing multimedia data. The study concludes by emphasizing the potential for wider applications of visual analytics in other sectors, suggesting future research directions to improve tool performance, scalability, and user accessibility, as well as exploring the integration of emerging technologies like artificial intelligence and virtual reality.

This study examines the influence of income inequality and poverty on the prevalence of stunting in ten provinces across Sumatra Island during the 2016–2024 period. Using a panel dataset of 90 observations and applying a Fixed Effect Model, the results indicate that both income inequality—measured by the Gini Ratio—and poverty have a positive and significant effect on stunting. The Gini Ratio shows a coefficient of 1.46 (p = 0.0002), while poverty records a coefficient of 6.28 (p = 0.0140), jointly explaining 52% of the variation in stunting prevalence. Spatial analysis further supports these findings, with Moran’s I values exceeding 0.40, suggesting strong spatial autocorrelation and clustering of high-stunting regions. High-risk clusters—Aceh, Jambi, and Bengkulu—are characterized by Gini Ratios above 0.33 and poverty levels exceeding 12%, reinforcing the existence of an intergenerational poverty–stunting trap, particularly influenced by urban–rural disparities (rural 53.3% vs urban 34.9%). The study highlights that specific nutrition interventions such as supplementary feeding, micronutrient programs, and breastfeeding promotion are insufficient without accompanying structural reforms addressing economic inequality. Therefore, multisectoral convergence strategies are required, including expanded conditional cash transfers, progressive local taxation reforms, nutrition-focused social assistance, and universal basic infrastructure to accelerate stunting reduction toward the 14.2% target by 2029.

The development of small and medium enterprises (SMEs) and household industries is often regarded as the economic foundation of a region. However, much of the existing research in Indonesia remains focused on quantitative descriptive analysis, while providing limited attention to spatial dynamics and interregional disparities. This study aims to critically evaluate the spatial distribution patterns of SMEs and household industries at the regency and city levels, with particular emphasis on clustering tendencies, unequal distribution, and their relationships with regional characteristics. A spatial analysis approach is employed to identify spatial autocorrelation and industrial clustering patterns, which is complemented by a structural analysis of infrastructure availability, market accessibility, and regional institutional capacity. The findings indicate that the distribution of SMEs and household industries is not geographically random, but rather forms clusters that are predominantly concentrated in areas with higher levels of accessibility and economic activity. This condition reflects spatial inequality that may exacerbate development disparities between regencies and cities. Furthermore, the results reveal that uniform industrial development policies are insufficient to accommodate the diverse spatial characteristics across regions. Therefore, this study underscores the importance of formulating spatially informed and context-sensitive policies for the development of SMEs and household industries in order to promote more balanced and sustainable regional industrial development.

This study aims to map the intellectual structure and research trends in entrepreneurial innovation using bibliometric analysis based on Scopus data. A total of 891 documents published between 1972-2025 were analyzed through Bibliometrix and Biblioshiny, employing techniques such as bibliographic coupling, co-authorship, and thematic mapping. The results reveal four major clusters: (1) innovation theory and entrepreneurial development, (2) business model innovation and digital transformation, (3) regional innovation systems and policy frameworks, and (4) sustainability and green entrepreneurship. Emerging themes include artificial intelligence (AI), generative AI applications, and digital entrepreneurship education, indicating a shift toward multi-level and interdisciplinary integration. Influential documents and authors were identified, highlighting their role in shaping the knowledge base. The findings suggest that entrepreneurial innovation research is evolving toward digitalization, sustainability, and policy-driven ecosystems, offering opportunities for longitudinal and mixed-method studies. This study contributes by providing a comprehensive overview of the field, identifying gaps, and proposing future research directions to strengthen theoretical and practical advancements.

Soil contamination by hydrocarbons, pesticides, heavy metals, and complex pollutants is rapidly increasing and degrading essential ecosystem functions. Physical or chemical treatments offer faster results, yet they are often costly, energy-intensive, and risk disrupting soil biological integrity without fully eliminating pollution sources. Microorganism-based bioremediation provides a more sustainable alternative by utilizing microbial metabolism to degrade or immobilize pollutants into less toxic and less mobile forms. This article presents a structured literature review on the roles and applications of microorganisms for bioremediation of contaminated soils, covering comparisons between single isolates and microbial consortia, dominant biological mechanisms, and ecological challenges in field application. A Systematic Literature Review approach was applied, using narrative synthesis and thematic clustering of national and international journals published between 2020 and 2025. The review indicates that single microbial isolates are commonly selected for specific pollutant targets, whereas microbial consortia are preferred for mixed or persistent contaminants due to metabolic synergy that enhances microbial adaptability and stepwise pollutant breakdown in highly polluted soils. Adaptive mechanisms such as EPS production and biofilm formation contribute to microbial resilience under stress and help retain contaminants within the soil matrix. Key challenges identified include inoculum stability under extreme conditions and limited microbial access to pollutants trapped in micro-soil pores. The findings highlight that microbial selection strategies must be tailored to pollutant characteristics and soil environmental conditions, while also emphasizing the potential of biofilm-based systems and organic carriers to support broader field implementation of microbial bioremediation.

Abstract: The rapid diffusion of Generative Artificial Intelligence (AI) has intensified public debate regarding its benefits, risks, and societal implications. This study investigates public sentiment and thematic structures surrounding Generative AI by analyzing Twitter discourse as a representation of large-scale, real-time public perception. The research addresses two main problems: how public sentiment toward Generative AI is distributed and what dominant themes shape this perception. Accordingly, the objective is to map both emotional polarity and thematic narratives embedded in social media conversations. A computational mixed-methods approach was employed using a dataset of 12,470 tweets collected on 17 December 2024. Sentiment classification was conducted using a transformer-based DistilBERT model, while semantic representations were generated with Sentence-BERT. Topic modeling was performed using BERTopic, integrating HDBSCAN clustering and class-based TF-IDF to extract coherent and interpretable topics. Human-in-the-loop validation supported the interpretive robustness of topic labeling. The findings reveal that public sentiment toward Generative AI is predominantly positive (41.8%), particularly in relation to productivity enhancement, education, and creative applications. Neutral sentiment (31.4%) reflects informational discourse, while negative sentiment (26.8%) centers on ethical concerns, privacy risks, misinformation, and AI hallucinations. Seven dominant topics were identified, with clear topic–sentiment alignment showing optimism in utility-driven themes and skepticism in ethics- and risk-related discussions. In conclusion, public perception of Generative AI is dualistic—characterized by strong enthusiasm alongside persistent caution. These results provide empirical insights for AI governance, responsible innovation, and future research on socio-technical impacts of Generative AI. *    

: The rapid expansion of cybersecurity standards and threat intelligence frameworks has led to significant semantic fragmentation among security terminologies, hindering effective information retrieval and interoperability across systems. Traditional keyword-based search approaches are inadequate for capturing the contextual meaning of security terms, particularly within formal frameworks such as NIST, MITRE ATT&CK, and CWE. This study addresses this challenge by proposing CyberBERT, a transformer-based semantic search framework designed to align cybersecurity terminologies through deep contextual representation and ontology-driven reasoning. Research Objectives: The primary objective of this research is to develop a semantic retrieval model capable of understanding conceptual relationships between security terms beyond lexical similarity. Methodology: The proposed methodology fine-tunes a BERT-based model on the NIST Glossary corpus using a combination of masked language modeling and triplet loss objectives to generate discriminative semantic embeddings. These embeddings are further aligned with cybersecurity ontologies, including MITRE ATT&CK and CWE, to enhance semantic consistency and explainability. Semantic retrieval is performed using cosine similarity within a 768-dimensional embedding space and evaluated using Mean Reciprocal Rank (MRR) and Precision@K metrics. Results: Experimental results demonstrate that CyberBERT achieves an MRR of 0.832, outperforming domain-adapted baselines such as SecureBERT and CyBERT. The integration of ontology alignment improves semantic accuracy by over 6%, while robustness evaluations confirm resilience against adversarial linguistic perturbations. Visualization using t-SNE reveals coherent semantic clustering aligned with the five core NIST Cybersecurity Framework functions. Conclusions: In conclusion, CyberBERT effectively bridges semantic gaps across cybersecurity terminologies by combining transformer-based contextual learning with ontological reasoning. The framework offers a robust, interpretable, and scalable solution for semantic search, supporting improved interoperability and knowledge discovery in cybersecurity operations and standards harmonization.

The disparity in the quality of rehabilitation services across regional work units presents a significant challenge to effective public management. This study aims to bridge the gap between problem diagnosis and policy prediction by proposing a hybrid, data-driven approach. We integrate K-Means Clustering to map the current state of service quality and Stochastic Simulation to predict the impact of strategic interventions. Using the 2024 Public Satisfaction Index (IKM) dataset from the National Narcotics Agency (BNN), the K-Means algorithm initially identified 26 work units (15.7%) in the "Red Zone" (critical performance), highlighting urgent areas for improvement. Next, a stochastic simulation modeling a "Directed Priority Intervention" scenario was run. The results predicted a significant structural shift in the distribution of service quality, characterized by an 80.8% decrease in critical units (down to 5 units) and a 71.8% increase in excellent performing units (up to 67 units). These findings validate that the integration of clustering and simulation provides a comprehensive framework for evidence-based decision-making, enabling policymakers to optimize resource allocation and efficiently accelerate national service standardization.

This study aims to examine customer segmentation through K-Means clustering from a customer data management perspective, emphasizing the interpretive value of analytical results rather than solely their computational outcomes. The research addresses a critical issue in contemporary data-driven organizations, where customer analytics is often reduced to technical modeling without sufficient translation into managerial insights. To respond to this gap, the study adopts a qualitative interpretive approach embedded within a quantitative clustering process, positioning clustering as part of a broader information management cycle. The empirical analysis is based on the Mall Customers Dataset obtained from Kaggle, consisting of 200 customer records with numerical attributes representing age, annual income, and spending score. Quantitative processing using K-Means clustering was employed to identify customer segments, while qualitative interpretation was applied to analyze the managerial meaning of each cluster. Data interpretation was supported by analytical documentation, visualization outputs, and reflective analysis of cluster characteristics. The findings reveal four distinct customer segments with different behavioral and economic profiles, each carrying specific strategic implications for customer relationship management and marketing decision-making. The study demonstrates that the primary value of clustering lies not merely in segment formation, but in its ability to transform raw customer data into actionable managerial knowledge. In conclusion, this research contributes to customer analytics literature by integrating data mining techniques with qualitative interpretation, offering a more human-centered and decision-oriented framework for customer data management. Future research is encouraged to extend this approach using organizational case studies or participatory decision-making contexts.

Village infrastructure development plays an important role in improving community welfare. However, public opinion regarding the condition of infrastructure is often not analyzed comprehensively and based on data. This study aims to analyze public opinions toward infrastructure in Malitidari Village using the Unsupervised Learning method with the Orange application. The data were collected from public comments on social media and digital surveys, which were then processed in text format. The analysis process includes data preprocessing, tokenization, vectorization, and the application of a clustering algorithm to group opinions into several categories without prior labeling. The Orange application was used due to its ability to visualize data analysis workflows interactively and efficiently. The results of this study show that public opinions can be grouped into several main clusters: positive opinions related to the development of roads and public facilities, and negative opinions concerning delays and unequal quality of infrastructure. Based on these findings, the Unsupervised Learning method is proven effective in illustrating public perceptions of village infrastructure conditions. The results are expected to serve as a reference for the village government in planning and improving the quality of infrastructure development in Malitidari Village  

Technological advancements in education have led to major transformations, particularly with the implementation of the Merdeka Curriculum, which emphasizes learning flexibility, student-centered approaches, and educator autonomy in developing innovative teaching methods. One of its essential aspects is the integration of technology for managing educational data, including student health records. At SMP IT Mutia Rahma, biannual student health monitoring has generated a growing volume of data, making it difficult to identify students experiencing psychological challenges. Adolescent mental health problems—such as learning stress, anxiety, and social pressure—can negatively affect academic performance if left unaddressed. This study aims to group students based on their mental health conditions to support more effective intervention strategies. The K-Means Algorithm, a data mining technique for clustering data by similarity, was employed to analyze student health data. The results show that in a three-cluster model, Cluster 2 represents students in a stable condition characterized by high resilience and low counseling needs, indicating good mental health and academic engagement. Meanwhile, Clusters 1 and 3 include students requiring further attention and support. This research demonstrates that the K-Means Algorithm can serve as an effective tool in identifying and categorizing student mental health conditions to improve school-based health management and early intervention programs.

The Merdeka Curriculum implemented in various schools in Indonesia aims to provide flexibility in learning, where students can learn according to their individual needs, interests, and pace. One of the challenges in implementing this curriculum is how to effectively identify classroom activity and student discipline. SD Islamiyah, as a school that implements the Merdeka Curriculum, also faces challenges in understanding variations in student classroom activity and discipline. Some students are able to learn in a disciplined manner with little guidance, while others require more intensive support from teachers. Therefore, a system is needed that can group student data more systematically so that teachers can develop teaching strategies that suit the needs of each group of students. One algorithm that can be used in data grouping is k-means clustering. The K-Means algorithm is a non-hierarchical algorithm derived from the data clustering method. The K-Means algorithm begins with the formation of cluster partitions at the beginning, then iteratively refines these cluster partitions until there are no significant changes in the cluster partitions. The K-Means method partitions data into groups so that data with similar characteristics are placed in the same group and data with different characteristics are grouped into other groups. This method can help group students more accurately based on their Class Activity and Discipline. From the results of the analysis, it was concluded that the student data group with Class Activity was Moderately Active Students, with Discipline being Disciplined, and an average score of 71-80.