Journal of Information Systems Engineering and Business Intelligence

Coastal tourism is a strategic sector in global economic development, but it faces serious challenges related to sustainability. Understanding tourist perceptions is crucial for maintaining destination competitiveness. Moreover, sentiment analysis based on online reviews offers new opportunities to explore tourist opinions broadly and in real-time. This study aims to conduct a Systematic Literature Review (SLR) of research discussing the application of sentiment analysis in coastal tourism from 2015 to 2025. The literature search was conducted in major databases, namely Scopus, ScienceDirect, IEEE Xplore, and ACM Digital Library, using predetermined search strings. Based on the initial search results obtained from 190 articles, after the duplication removal and Quality Assessment (QA) processes, only 36 articles met the criteria and were further analyzed. The results demonstrate an increasing trend in publications since 2018, peaking in 2024. These studies are spread across various regions of the world, with a predominance of Southeast Asia and Europe. TripAdvisor and Google Maps served as the primary data sources. Analysis methods included lexicon approaches, Machine Learning (ML) algorithms (Naive Bayes, Support Vector Machine (SVM), k-Nearest Neighbor (kNN), Random Forest), Deep Learning (DL) (Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM)), as well as hybrid approaches and Large Language Models (LLM). The labeling schemes used varied, ranging from binary and three-class to emotional categories, satisfaction dimensions, and the Kano model. Overall, tourist reviews of coastal destinations were positive, although complaints about cleanliness, price, service, and sustainability were common. These findings confirm that sentiment analysis is a crucial tool for understanding tourist experiences and can support sustainable coastal tourism management globally.

Clickbait headlines are widely used in online media to attract readers through exaggerated or misleading titles, potentially leading to user dissatisfaction and information overload. This study proposes a machine learning approach for detecting clickbait in Indonesian news headlines using classical classification models and ensemble learning. The dataset consists of labeled clickbait and non-clickbait headlines in Bahasa Indonesia, which were processed and represented using TF-IDF vectorization. Three base classifiers, Multinomial Naive Bayes, Logistic Regression, and Support Vector Machine, were integrated using soft voting and stacking ensemble methods. The experimental results indicate that the stacking ensemble model achieved the highest accuracy of 0.7728, while the voting ensemble recorded the best F1-score of 0.7080, outperforming individual classifiers. Despite these gains, the SVM model demonstrated the most substantial decline in accuracy after stopwords removal, dropping by 0.0410. These findings highlight the effectiveness of ensemble learning in enhancing clickbait detection performance and suggest potential for further optimization in model selection and integration strategies.

A graph G admits an H-covering if every edge of G belongs to a subgraph isomorphic to a given graph H. G is said to be H-magic if there exists a bijection $f:V\left(G\right)\cup E\left(G\right)\to \{1,2,\dots ,|V\left(G\right)|+|E\left(G\right)\left|\right\}$ such that $w_f\left(H^{\prime }\right)={\sum }_{v\in V\left(H^{\prime }\right)}f\left(v\right)+{\sum }_{e\in E\left(H^{\prime }\right)}f\left(e\right)$ is a constant, for every subgraph $H^{\prime }$ isomorphic to H. In particular, G is said to be H-supermagic if $f\left(V\right(G\left)\right)=\{1,2,\dots ,|V\left(G\right)\left|\right\}$. When H is isomorphic to a complete graph $K_2$, an H-(super)magic labeling is an edge-(super)magic labeling. Suppose that G admits an F-covering and H-covering for two given graphs F and H. We define G to be $(F,H)$-sim-(super)magic if there exists a bijection $f^{\prime }$ that is simultaneously F-(super)magic and H-(super)magic. In this paper, we consider $(K_2,H)$-sim-(super)magic where H is isomorphic to three classes of graphs with varied symmetry: a cycle which is symmetric (both vertex-transitive and edge-transitive), a star which is edge-transitive but not vertex-transitive, and a path which is neither vertex-transitive nor edge-transitive. We discover forbidden subgraphs for the existence of $(K_2,H)$-sim-(super)magic graphs and classify classes of $(K_2,H)$-sim-(super)magic graphs. We also derive sufficient conditions for edge-(super)magic graphs to be $(K_2,H)$-sim-(super)magic and utilize such conditions to characterize some $(K_2,H)$-sim-(super)magic graphs.

On (F, H)-simultaneously-magic labelings of graphs

Software defect prediction (SDP) often faces challenges related to heterogeneous software metrics, classifier dependency, and severe class imbalance, which may limit the robustness and generalization of feature selection strategies. This study proposes an adaptive feature selection approach to construct structured and discriminative feature subsets that remain effective across diverse datasets and learning models. The proposed method first measures the relationship between each software metric and the defect label using absolute determination power and then applies an adaptive retention rule to iteratively retain features with stronger discriminative contribution. The evaluation was conducted on multiple public defect datasets using several classical machine learning classifiers. Unlike approaches optimized for specific classifiers, the proposed strategy emphasizes cross-classifier robustness and imbalance-aware evaluation through defect recall and Matthews correlation coefficient. Experimental results show that the proposed method achieves a competitive average MCC of 0.257 and a defect recall of 0.446 compared with baseline approaches, although the statistical tests do not indicate significant superiority. Therefore, the proposed method should be interpreted as a comparable and stable alternative for feature selection under imbalanced SDP conditions. Stability and statistical analyses further indicate that the proposed method maintains comparable performance across dataset-classifier combinations. In addition, feature compactness analysis shows that the performance gains are attributable to efficient, interpretable feature subsets, highlighting the importance of robustness-oriented feature selection in SDP.