Academic stress is a common psychological problem experienced by students during their education at university. This condition can be influenced by various factors, including high academic load, limited time to complete assignments, difficulty understanding lecture material, and an unsupportive learning environment. If not managed effectively, academic stress has the potential to negatively impact students' academic achievement, learning motivation, and mental health. Along with the development of information technology, the application of data mining techniques can be utilized to help identify and classify students' academic stress levels more objectively and accurately. This study aims to classify the academic stress levels of using the C4.5 algorithm. Research data were obtained through distributing questionnaires to students from the Informatics Engineering, Information Systems, and Computer Engineering Study Programs. The research stages include data preprocessing, the formation of a classification model, and evaluation of model performance. Algorithm was chosen because it has the ability to build a classification model in the form of a decision tree that is easy to interpret and is able to determine the most influential attributes in the decision-making process. The resulting classification model is expected to be a tool for universities, especially lecturers and student affairs, in conducting early detection of students who have the potential to experience academic stress. Thus, appropriate mentoring and intervention steps can be provided to support the academic success and psychological well-being of students. Keywords: Academic Stress ,C4.5 Algorithm, Classification,Data Mining, Students
This study analyzes the polarization of public opinion regarding the removal of fuel subsidies in Indonesia using a graph-based text classification approch that integrates Pointwise Mutual Information (PMI) and Graph Convolution Networks (GCN). One of the main challenges in sentiment analysis of social media data lies in capturing contextual and semantic relationships within short and noisy texts, which often limits the performance of conventional classifications methods. A dataset consisting of 602 tweets was preprocessed and represented as a graph, where word-word relationships were weighted using PMI and document-word connections were weighted using TF-IDF. Two model configurations were evaluated: a baseline GCN without PMI and a PMI-enhanced GCN, both trained using a 70:15:15 train-validation-test split with indentical hyperparameter settings. Experimental results indicate that the baseline GCN achieved a maximum accuracy of 65%, while the integration of PMI improved performance, yielding the highest the accuracy of 73%. These results demonstrate that incorporating PMI enriches graph representations and enhances the effectiveness of GCN in sentiment classification of political opinion data from social media.
Perkembangan persaingan usaha yang semakin kompetitif menuntut pelaku bisnis untuk menerapkan strategi yang mampu mempertahankan dan meningkatkan loyalitas pelanggan. Salah satu strategi yang dapat diterapkan adalah Customer Relationship Management (CRM), yang berfokus pada pengelolaan hubungan jangka panjang dengan pelanggan. Penelitian ini bertujuan untuk menganalisis peran CRM dalam meningkatkan loyalitas pelanggan di Toko Fasya Pane. Hasil penelitian menunjukkan bahwa penerapan CRM memberikan kemudahan bagi pelanggan dalam melakukan pembelian barang, meningkatkan kualitas pelayanan, serta menciptakan pengalaman berbelanja yang lebih efektif dan nyaman. Pelayanan yang responsif, komunikasi yang baik, serta perhatian terhadap kebutuhan pelanggan terbukti mampu meningkatkan tingkat kepuasan pelanggan. Peningkatan kepuasan tersebut berdampak positif terhadap terbentuknya loyalitas pelanggan yang berkelanjutan. Dengan demikian, penerapan CRM berperan penting dalam memperkuat hubungan antara Toko Fasya Pane dan pelanggan serta mendukung keberlangsungan usaha secara jangka panjang.
Rainfall prediction is an important component of weather analysis as it provides valuable information to support decision-making in sectors such as agriculture, transportation, and environmental management. Although various studies have compared machine learning algorithms for rainfall classification, many of them lack detailed discussion on dataset characteristics and practical system implementation. Therefore, this study aims to evaluate and compare the performance of Decision Tree and Naive Bayes algorithms for rainfall classification while considering dataset characteristics and implementing the model in a web-based application. The dataset used in this study consists of 2,500 records with meteorological parameters including temperature, humidity, wind speed, cloud cover, and atmospheric pressure. The data underwent preprocessing, including data cleaning and label encoding, where rainfall was represented as 1 and no rainfall as 0. The dataset was divided into training and testing sets, and both algorithms were applied to build classification models. Model performance was evaluated using confusion matrix, accuracy, and ROC curve analysis. The results show that the Decision Tree algorithm achieved an accuracy of 1.00 (100%), while the Naive Bayes algorithm achieved 0.972 (97.2%). Although Decision Tree shows superior performance, the perfect accuracy may indicate potential overfitting, and therefore the results should be interpreted carefully. Furthermore, the developed models were successfully implemented into a web-based application that enables users to perform rainfall prediction interactively. This study demonstrates that Decision Tree provides better performance for rainfall classification in the given dataset, while also highlighting the importance of considering dataset characteristics and evaluation methods. The integration of machine learning models into a web-based system provides a practical contribution for real-world weather prediction applications. Keywords: Rainfall Classification, Decision Tree, Naive Bayes, Machine Learning, Weather Dataset, Web-Based Application
Breast cancer remains one of the most prevalent causes of cancer-related mortality among women worldwide, making early and accurate detection critically important. Machine learning techniques have been widely applied for this purpose; however, many existing studies primarily focus on predictive accuracy without providing comprehensive analysis of model optimization and interpretability. This study proposes a comparative framework integrating Support Vector Machine (SVM) and Extreme Gradient Boosting (XGBoost) with Bayesian Optimization to enhance hyperparameter tuning and model performance. The Breast Cancer Wisconsin Dataset, consisting of 569 samples with 30 numerical features, is used for evaluation. The proposed approach includes data preprocessing, dataset splitting, systematic hyperparameter optimization, model training, and performance evaluation. Experimental results show that the XGBoost model achieves superior performance compared to SVM, with an accuracy of 98.24% and an Area Under the Curve (AUC) of 0.994. Further analysis indicates that the model maintains a strong balance between precision and recall, with minimal misclassification. In addition, feature importance analysis reveals that attributes related to tumor size and structural irregularities contribute significantly to the prediction results, supporting the interpretability of the model in a medical context. The main contribution of this study lies in providing a more comprehensive evaluation that combines performance comparison, optimization effectiveness, and feature-level interpretation within a unified framework. The findings demonstrate that the integration of XGBoost and Bayesian Optimization offers a reliable and interpretable approach for breast cancer classification, with strong potential for implementation in machine learning–based clinical decision support systems. Keywords: breast cancer, machine learning, XGBoost, Support Vector Machine, Bayesian Optimization.
