Publication Search

-

Student dropout in higher education remains a persistent socioeconomic challenge, yet many predictive models reported in the literature are methodologically compromised by randomized cross-validation schemes that introduce temporal data leakage and artificially inflate predictive performance. This study proposes a longitudinal prescriptive learning analytics framework integrating three complementary methodological components: a Leave-One-Cohort-Out (LOCO) temporal validation protocol, a hybrid SMOTE-ENN class balancing strategy, and temporal velocity feature engineering derived from Learning Management System (LMS) behavioral trajectories. The framework was evaluated on a longitudinal dataset comprising 464,739 enrollment records and 77 features. Five predictive algorithms—XGBoost, LightGBM, CatBoost, Random Forest, and Logistic Regression—were comparatively assessed on a strictly isolated blind holdout cohort (2022), with CatBoost emerging as the champion estimator, achieving a PR-AUC of 0.8859, a Macro F1-Score of 0.9143, and the lowest Brier Score (0.0221), thereby demonstrating superior calibration and discriminative capability under severe class imbalance (93:7 ratio). Comprehensive ablation analysis revealed that temporal velocity features function not merely as additive predictors, but as a structural prerequisite enabling Synthetic Minority Oversampling Technique with Edited Nearest Neighbors (SMOTE-ENN) to generate high-quality synthetic boundary instances; removing these features reduced minority-class precision from 0.8302 to 0.6721. To operationalize predictive outputs into actionable intervention pathways, Diverse Counterfactual Explanations (DiCE) were implemented under a three-tier causal constraint architecture on 96 borderline high-risk students, generating 384 feasible intervention scenarios exclusively targeting forward-looking behavioral velocity metrics without constraint violations. Collectively, these findings advance the paradigm of prescriptive learning analytics by providing educational institutions with interpretable risk diagnostics and operationally feasible intervention guidance grounded in empirically validated behavioral and temporal dynamics.

This research aims to classify students’ lifestyles using the K-Nearest Neighbors (KNN) algorithm. The dataset consists of 392 high school students obtained from Kaggle, with key attributes including study hours, social media usage, Netflix viewing duration, attendance, sleep quality, internet quality, mental health, and extracurricular activities. KNN was chosen for its simplicity in distance-based classification, measured using Euclidean Distance. The data was divided into training and testing sets, then evaluated using accuracy and a confusion matrix. The results show that KNN effectively classifies students’ lifestyles into four categories: healthy, less active, at risk, and highly at risk. This classification is expected to assist educational institutions, parents, and students in understanding lifestyle patterns and their impact on academic performance and mental well-being. Furthermore, this study emphasizes the relevance of applying machine learning in education, aligned with Islamic values concerning health, discipline, and the optimal use of time.

This research aims to classify students’ lifestyles using the K-Nearest Neighbors (KNN) algorithm. The dataset consists of 392 high school students obtained from Kaggle, with key attributes including study hours, social media usage, Netflix viewing duration, attendance, sleep quality, internet quality, mental health, and extracurricular activities. KNN was chosen for its simplicity in distance-based classification, measured using Euclidean Distance. The data was divided into training and testing sets, then evaluated using accuracy and a confusion matrix. The results show that KNN effectively classifies students’ lifestyles into four categories: healthy, less active, at risk, and highly at risk. This classification is expected to assist educational institutions, parents, and students in understanding lifestyle patterns and their impact on academic performance and mental well-being. Furthermore, this study emphasizes the relevance of applying machine learning in education, aligned with Islamic values concerning health, discipline, and the optimal use of time.

This study presents a comprehensive comparative analysis of four traditional machine learning algorithms Decision Tree, Random Forest, K-Nearest Neighbors, and Support Vector Machine for Android malware detection using the preprocessed TUANDROMD dataset comprising 4,465 instances and 241 features representing both static and dynamic application characteristics. Motivated by the limitations of conventional signature-based and hybrid detection methods, especially in managing imbalanced datasets and detecting emerging malware variants, the study employed SMOTE to ensure balanced training data and fair model evaluation. The dataset was divided into 80% training and 20% testing subsets, and models were assessed using key performance metrics including accuracy, precision, recall, F1-score, and ROC AUC. The findings revealed that the proposed Random Forest model outperformed the other classifiers, achieving an accuracy of 0.993, precision of 0.992, recall of 1.000, F1-score of 0.996, and a near-perfect ROC AUC of 0.9998 surpassing state-of-the-art approaches. These results affirm the superior predictive capability, consistency, and robustness of the Random Forest algorithm in Android malware detection. The study concludes that base models, when integrated with class-balancing techniques, provide reliable and efficient malware detection across imbalanced datasets. For future research, the study recommends exploring advanced hybrid or ensemble frameworks that integrate Random Forest with deep learning architectures or other meta-heuristic optimization techniques to further enhance detection accuracy, adaptability, and resilience against rapidly evolving Android malware threats.

Detection and identification of plant diseases is critical to the success and efficiency of agricultural production. Plant disease outbreaks are becoming more frequent throughout the world, and the presence of these diseases in cultivated plants has a significant impact on productivity. Therefore, researchers are focusing on developing effective and reliable plant disease detection methods. Thus, farmers can take advantage of early detection of this disease to minimize future losses. This article discusses machine learning approaches as well as decision trees, K-nearest neighbors, naive Bayes, support vector machines (SVM), and random forests for detecting coffee leaf diseases using leaf images. The above-mentioned classifications were researched and compared to determine the most suitable plant disease prediction model with the highest accuracy. Compared with other classification algorithms, the SVM algorithm achieves the highest accuracy of 99.75%. All the models trained above will be used by farmers to quickly identify and classify new diseases in images as a prevention strategy. As a preventive measure, farmers can detect and classify new diseases in images early.

Aritmia merupakan gangguan irama jantung yang umum terjadi pada lansia dan dapat menimbulkan risiko kesehatan serius jika tidak terdeteksi secara dini. Penelitian yang dilakukan bertujuan untuk mengidentifikasi aritmia pada lansia menggunakan algortima K- Nearest Neighbor (KNN) berdasarkan data elektrokardiogram (EKG). Data yang digunakan berjumlah 105 data EKG lansia yang diperoleh dalam format CSV. Proses awal melibatkan pembersihan dan normalisasi data menggunakan metode StandardScaler, serta pelabelan awal menggunakan algoritma K-Means Clustering untuk mengelompokkan data ke dalam dua kelas: Normal dan Sangat Berpotensi Aritmia. Data kemudian dibagi menjadi 70% data latih dan 30% data uji dengan metode stratified split untuk menjaga proporsi label. Model KNN dilatih dengan parameter k = 3, dan dievaluasi menggunakan confusion matrix serta classification report. Hasil pengujian menunjukkan akurasi model sebesar 97% dengan nilai precision dan recall yang tinggi pada kedua kelas. Hasil ini menunjukkan bahwa algoritma KNN efektif dalam mengklasifikasikan kondisi aritmia pada lansia dan memiliki potensi untuk diterapkan dalam sistem pendukung diagnosis berbasis data EKG.

Student dropout remains a significant challenge for higher education institutions as it impacts academic quality, educational management efficiency, and students' success in completing their studies. Therefore, an approach that can identify students at risk of dropping out is necessary so that timely academic interventions can be made. This study aims to develop a dropout detection model using an Artificial Neural Network (ANN). The data used come from a publicly available higher education dataset, ensuring research reproducibility. Data preprocessing steps were carried out to improve data quality before modeling, and the Synthetic Minority Over-Sampling Technique combined with Edited Nearest Neighbors (SMOTE-ENN) was applied to address class imbalance issues. The ANN model's performance was evaluated using accuracy, precision, recall, F1-score, and area under the ROC curve (ROC-AUC). The test results show that the ANN model can provide excellent predictive performance in detecting at-risk students. The application of SMOTE-ENN also proved to enhance the model’s sensitivity toward the minority class, as indicated by improvements in recall and F1-score. These findings indicate that the developed ANN model has the potential to be used as a student dropout detection system to support data-driven decision-making and strategy development within higher education institutions.

AImbalanced data remains a significant issue in heart disease classification using machine learning, as it tends to cause models to overestimate the majority class while ignoring minority classes with high clinical value. This can lead to a decrease in accuracy and the model's ability to accurately detect disease cases. Therefore, this study aims to assess the effectiveness of oversampling techniques, namely Random Oversampling and Synthetic Minority Oversampling Technique (SMOTE), in improving the performance of the K-Nearest Neighbors (KNN), Naive Bayes (NB), and Random Forest (RF) algorithms. The dataset used comes from Kaggle and consists of 918 data sets with 12 attributes representing patient information related to heart disease prediction. The research stages include data preprocessing, baseline model testing, and re-evaluation using the two oversampling methods. Experimental results show that oversampling can improve the performance of all algorithms. KNN achieved the best results with SMOTE, with an accuracy of 72.98% and an F1-score of 75.39%. In the Naive Bayes algorithm, both oversampling techniques produced relatively stable performance, with the highest F1-score of 73.56% using SMOTE. Meanwhile, Random Forest showed the most optimal performance when combined with Random Oversampling, with an accuracy of 79.19% and an F1-score of 81.51%. These findings confirm that the success of data balancing techniques is strongly influenced by the characteristics of the classification algorithm used, and provide a practical contribution in determining strategies for handling imbalanced data in health research.

Manual quality assessment of Platelet Concentrate (TC) is highly subjective and inconsistent, necessitating an objective, automated classification system. This study aims to develop a computationally efficient, low-cost model for TC quality classification using Histogram Features extracted from grayscale images combined with the K-Nearest Neighbor (KNN) algorithm. The methodology employed critical preprocessing steps, including StandardScaler for normalization and SMOTE for balancing the training data, followed by optimization across K=1 to K=30. The optimal model achieved a maximum accuracy of 69.23% at K=6, with an F1-Score of 71.43%, confirming robust performance on the imbalanced testing set. The results validate the effectiveness of the Histogram-KNN approach as a consistent and reliable decision support system for rapid TC quality screening in resource-limited settings.

The world of work is an environment related to the work we are currently in. In other words, it is a place where various individuals perform an activity. The quality of college graduates is not only seen in terms of high or good grades / GPA. There are many other considerations, where large companies see a potential possessed by the person concerned. The dataset in this study was taken from student respondents about the world of work. One way to classify the influence of competence on the world of work in machine learning is to use datasets as training data so that performance testing can be carried out with the right classification method. From the results of the tests carried out, it is concluded that the results of the comparison are different, which shows the accuracy value of KNN which is around 96%, while the results of the SVM accuracy tested are 98%, so that the accuracy of SVM is better than KNN.

The rapid growth of open-source software (OSS) in machine learning (ML) has intensified the need for reliable, automated methods to assess project quality, particularly as OSS increasingly underpins critical applications in science, industry, and public infrastructure. This study evaluates the effectiveness of a diverse set of machine learning and deep learning (ML/DL) algorithms for classifying GitHub OSS ML projects as engineered or non-engineered using a SMOTE-enhanced and explainable modeling pipeline. The dataset used in this research includes both numerical and categorical attributes representing documentation, testing, architecture, community engagement, popularity, and repository activity. After handling missing values, standardizing numerical features, encoding categorical variables, and addressing the inherent class imbalance using the Synthetic Minority Oversampling Technique (SMOTE), seven different classifiers—K-Nearest Neighbors (KNN), Decision Tree (DT), Random Forest (RF), XGBoost (XGB), Logistic Regression (LR), Support Vector Machine (SVM), and a Deep Neural Network (DNN)—were trained and evaluated. Results show that LR (84%) and DNN (85%) outperform all other models, indicating that both linear and moderately deep non-linear architectures can effectively capture key quality indicators in OSS ML projects. Additional explainability analysis using SHAP reveals consistent feature importance across models, with documentation quality, unit testing practices, architectural clarity, and repository dynamics emerging as the strongest predictors. These findings demonstrate that automated, explainable ML/DL-based quality assessment is both feasible and effective, offering a practical pathway for improving OSS sustainability, guiding contributor decisions, and enhancing trust in ML-based systems that depend on open-source components.

Anxiety disorders are common mental health problems in society, often unrecognized by the sufferer. Identifying the type of anxiety disorder and its influencing factors is crucial for proper treatment. This research aims to apply the K-Nearest Neighbor (K-NN) method in identifying types of anxiety disorders based on influencing factors, focusing on patient data from Sylvani Hospital, Binjai. The K-NN method was chosen because of its ability to classify based on data proximity. This study used medical record data of patients with anxiety disorders, which were processed using MATLAB and Microsoft Excel software. The results show that the K-NN method is effective in identifying types of anxiety disorders, with a high level of accuracy, especially in the identification of Panic Disorder (K05) and Social Anxiety Disorder (K03). The use of MATLAB simplified the identification process by automating results, while data processing in Excel improved classification accuracy. This study concludes that the K-NN method can be an effective alternative in identifying anxiety disorder types based on the factors that influence them. It is recommended for future research to involve more variables and mental health experts for a more comprehensive validation of the results.

Speech delay in children is a developmental issue commonly encountered in society, which can affect various aspects of a child's life, including communication, social interaction, and academic development. Early detection of speech delay is crucial for providing appropriate interventions to minimize its long-term impact on the child. This study aims to introduce the use of machine learning algorithms in detecting speech delay symptoms in children. Three machine learning algorithms applied in this study are Naïve Bayes, C4.5, and K-Nearest Neighbor (K-NN). These algorithms are used to classify speech delay symptoms based on health data, medical history, and environmental factors such as speaking habits and eating patterns. The outreach was conducted at Puskesmas Kota Rantauprapat with the involvement of parents and healthcare providers as participants. The experimental results showed that all three algorithms performed well in terms of accuracy, though with varying error rates. Naïve Bayes achieved relatively high accuracy but had a higher false positive rate compared to C4.5 and K-NN. C4.5 provided more stable results and was easier to interpret due to its decision tree structure. Meanwhile, K-NN performed better with data that had irregular distribution. This outreach is expected to assist both the community and healthcare providers in early detection of speech delay in children, providing a more efficient and affordable means for early intervention, which ultimately leads to better outcomes for children with speech delay.

One approach the government employs to decorate public welfare, mainly among low-income families, is through social help initiatives. however, the subjectivity inside the choice process regularly ends in mistargeting all through implementation. This observe objectives to apply the ok-Nearest Neighbor (ok-NN) and Naive Bayes algorithms inside a decision support device to perceive eligible recipients based on community statistics. The ok-NN algorithm determines similarity by calculating the Euclidean distance among new and current facts, whilst the Naive Bayes set of rules utilizes a probabilistic method based at the likelihood of attribute incidence inside each elegance. Key criteria considered consist of household income, employment kind, number of dependents, housing conditions, and asset possession. Experimental consequences reveal that each algorithms are powerful in as it should be classifying eligibility for help, with k-NN barely outperforming Naive Bayes. therefore, the combination of these algorithms can support stakeholders in making extra goal and efficient selections regarding the distribution of social useful resource.

Human Resource Management (HRM) plays a strategic role in improving organizational competitiveness through proper management of employee placement, training, and performance evaluation. To support the achievement of these goals, a predictive model is needed that can provide an accurate picture of employee performance. This study utilizes a Human Resource Management (HRM) dataset of 1,200 data and applies several classification algorithms to compare their effectiveness, namely J48 or C4.5, Random Forest, Naive Bayes, K-Nearest Neighbor (KNN), Logistic Regression, and Support Vector Machine (SVM). To obtain more optimal results, this study uses resampling techniques and attribute selection methods with a correlation attribute eval approach, so that class distribution can be more balanced and model accuracy increases. From the test results, the Decision Tree J48 algorithm showed the best performance with an accuracy level reaching 95.41%, a kappa value of 0.8925, a mean absolute error (MAE) of 0.0432, a precision of 0.955, a recall of 0.954, and an area under the ROC curve of 0.964. These findings indicate that J48 has excellent predictive capabilities compared to other algorithms. Furthermore, this study also found that the most influential variables in determining employee performance include the percentage of the last salary increase (EmpLast Salary Hike Percent), the level of work environment satisfaction (Emp Environment Satisfaction), the length of time since the last promotion (Years Since Last Promotion), and experience in the current role (Experience Years in Current Role). Overall, the results of the study indicate that the C4.5 algorithm with the application of the resampling technique can be an optimal solution in building an employee performance prediction system. Thus, this model has the potential to be a strong basis for managerial decision-making, particularly in designing HR development strategies and policies to improve organizational performance.

This study aims to utilize data mining techniques, specifically the K-Nearest Neighbors (KNN) algorithm, to classify leaf diseases in sugarcane (Saccharum officinarum). Early and accurate detection of leaf disease types is a crucial step in prevention and control strategies, thereby reducing potential crop losses caused by pathogen attacks. Leaf diseases in sugarcane, such as leaf scald, rust, and mosaic virus, are known to affect photosynthesis, inhibit growth, and reduce the quality and quantity of sugarcane produced. The classification process in this study was carried out through image analysis of infected sugarcane leaves, where features such as color, texture, and shape were extracted using digital image processing techniques. The KNN algorithm was chosen because of its non-parametric nature, ease of implementation, and its ability to provide accurate classification results even with limited data size. The working principle of KNN is to determine the class of a new sample based on the majority class of its k nearest neighbors in the feature space, making it very suitable for the case of leaf disease image classification. In addition to building a classification model, this study also examines disease prevention strategies based on the identification results. These strategies include the use of disease-resistant sugarcane varieties, the implementation of appropriate planting patterns, land moisture management, regular plantation sanitation, and the measured and environmentally friendly use of pesticides or fungicides. Model performance evaluation was conducted using accuracy, precision, recall, and F1-score metrics to assess model effectiveness across various data scenarios. The results of this study are expected to not only contribute to the development of decision support systems for farmers and related parties but also support the application of artificial intelligence-based technology in the agricultural sector.

Electricity has emerged as an essential requirement in modern life. As demand escalates, electricity costs rise, making wastefulness a drain on financial resources. Consequently, forecasting electricity usage can enhance our management of consumption. This study presents an IoT-based monitoring and forecasting system for electricity consumption. The system comprises two NodeMCU micro-controllers, a PZEM-004T sensor for collecting real-time power data, and three relays that regulate the current flow to three distinct electrical appliances. The data gathered is transmitted to a web application utilizing the k-Nearest Neighbor (k-NN) algorithm to forecast future electricity usage based on historical patterns. We evaluated the system's performance using four weeks of electricity consumption data. The results indicated that predictions were most accurate when the user’s daily consumption pattern remained stable, achieving a Mean Absolute Error (MAE) of approximately 1 watt and a Mean Absolute Percentage Error (MAPE) ranging from 1% to 1.7%. Additionally, predictions were notably precise during the early morning hours (3:00 AM to 8:00 AM) when k=6 was employed. This study demonstrates the effectiveness of integrating IoT-based systems with machine learning for real-time energy monitoring and forecasting. Furthermore, it emphasizes the application of data mining techniques within embedded IoT environments, providing valuable insights into the implementation of lightweight machine learning for smart energy systems.

This study aims to analyze and compare the performance of the Naive Bayes, K-Nearest Neighbors (KNN), and Decision Tree algorithms in predicting the purchase intention of e-commerce visitors using the Online Shoppers Purchasing Intention Dataset, which consists of 12,330 records and 18 variables, with the Revenue variable serving as the classification target. The preprocessing stage involved transforming categorical and boolean variables into numerical form, standardizing features using StandardScaler, and splitting the dataset into 80 percent training data and 20 percent testing data. Model evaluation was conducted using accuracy, precision, recall, F1-score, and ROC-AUC metrics, and was further strengthened by 10-fold cross-validation to obtain more stable results. The findings indicate that KNN achieved the highest accuracy of 0.866180, while Naive Bayes produced the highest recall value of 0.690998 and the highest ROC-AUC value of 0.821696. Meanwhile, Decision Tree demonstrated relatively balanced performance with an accuracy of 0.857259 and an F1-score of 0.571776, whereas the cross-validation results identified KNN as the model with the highest average accuracy of 0.8770. These findings suggest that the selection of a classification model for purchase intention prediction cannot rely solely on a single evaluation metric, as each algorithm possesses different strengths. Therefore, a comparative approach among algorithms can help determine the most suitable model for supporting consumer behavior analysis on e-commerce platforms.

The issue of determining the number of students' graduation times is one of the important indicators in transmitting the quality and effectiveness of the higher education process in universities. The rate of on-time graduation not only impacts accredited institutions, but also becomes a concern for campus management in designing learning strategies and academic guidance. This study aims to apply and compare two classification algorithms in data mining, namely C4.5 and K-Nearest Neighbor KNN, in predicting the accuracy of students' graduation times. Predictions are made based on academic attributes such as Grade Point Average GPA, number of credits that have been achieved, and Semester Grade Point Average IPS as input variables. The method used in this study is Knowledge Discovery in Database KDD which includes data selection, preprocessing, transformation, data mining, and evaluation of results. The study was conducted using the RapidMiner tool, with a dataset of 279 Informatics Study Program students from the 2015 to 2019 intake. The data was classified into two categories: "graduated on time" and "not graduated on time". The test results showed that the KNN algorithm provided better performance compared to C4.5. KNN produced an accuracy of 76.08%, with a precision of 73.11% and a recall of 41.92%. Meanwhile, the C4.5 algorithm produced an accuracy of 73.49%, with a precision of 64.62% and a recall of 41.89%. This difference in accuracy indicates that KNN is more effective in capturing patterns in the data and providing more accurate predictions in this context. Thus, the KNN algorithm can be considered a more optimal method to assist universities in predicting potential student admissions in a timely manner, thus enabling early intervention for students at risk of late graduation. This research also contributes to the development of data mining-based academic decision support systems in higher education.