Lung cancer is a leading cause of mortality worldwide, and early detection is crucial in improving treatment outcomes and reducing death rates. However, diagnosing medical images, such as Computed Tomography scans (CT scans), is complex and requires a high level of expertise. This study focuses on developing and evaluating the performance of Convolutional Neural Network (CNN) models, specifically the Visual Geometry Group 16 (VGG16) architecture, to classify lung cancer CT scan images into three categories: Normal, Benign, and Malignant. The dataset used consists of 1097 CT images from 110 patients, categorized according to these severity levels. The research methodology began with data collection and preparation, followed by training and testing the VGG16 model and comparing its performance with other CNN architectures, including Residual Network with 50 layers (ResNet50), Inception Version 3 (InceptionV3), and Mobile Neural Network Version 2 (MobileNetV2). The experimental results indicate that VGG16 achieved the highest classification performance, with a Test Accuracy of 98.18%, surpassing the other models. This accuracy highlights VGG16’s strong potential as a supportive diagnostic tool in medical imaging. However, a limitation of this study is the dataset size, which may reduce model accuracy when applied to new data. Future studies should consider increasing the dataset size, using Data Augmentation techniques, fine-tuning model parameters, and employing advanced models such as 3D CNN or Vision Transformers. Additionally, incorporating Gradient-weighted Class Activation Mapping (Grad-CAM) to interpret model decisions would enhance transparency and reliability. This study confirms the potential of CNNs, particularly VGG16, for classifying lung cancer CT images and provides a foundation for further development in medical applications.

Background: Class imbalance and limited interpretability remain major barriers to the clinical adoption of machine learning in diabetes prediction. These challenges often result in poor sensitivity to high-risk cases and reduced trust in AI-based decision support. This study addresses these limitations by integrating SMOTE-based resampling with SHAP-driven explainability, aiming to enhance both predictive performance and clinical transparency for real-world deployment. Objective: To develop and validate an interpretable machine learning framework that addresses class imbalance through advanced resampling techniques while providing clinically meaningful explanations for enhanced decision support. This study serves as a methodologically rigorous proof-of-concept, prioritizing analytical integrity over scale. While based on a computationally feasible subset of 1500 records, future work will extend to the full 100,000-patient dataset to evaluate scalability and external validity. We used the publicly available, de-identified Diabetes Prediction Dataset hosted on Kaggle, which is synthetic/derivative and not a clinically curated cohort. Accordingly, this study is framed as a methodological proof-of-concept rather than a clinically generalizable evaluation. Methods: We implemented a robust seven-stage pipeline integrating the Synthetic Minority Oversampling Technique (SMOTE) with SHapley Additive exPlanations (SHAP) to enhance model interpretability and address class imbalance. Five machine learning algorithms—Random Forest, Gradient Boosting, Support Vector Machine (SVM), Logistic Regression, and XGBoost—were comparatively evaluated on a stratified random sample of 1500 patient records drawn from the publicly available Diabetes Prediction Dataset (n = 100,000) hosted on Kaggle. To ensure methodological rigor and prevent data leakage, all preprocessing steps—including SMOTE application—were performed within the training folds of a 5-fold stratified cross-validation framework, preserving the original class distribution in each fold. Model performance was assessed using accuracy, area under the receiver operating characteristic curve (AUC), sensitivity, specificity, F1-score, and precision. Statistical significance was determined using McNemar’s test, with p-values adjusted via the Bonferroni correction to control for multiple comparisons. Results: The Random Forest-SMOTE model achieved superior performance with 96.91% accuracy (95% CI: 95.4–98.2%), AUC of 0.998, sensitivity of 99.5%, and specificity of 97.3%, significantly outperforming recent benchmarks (p < 0.001). SHAP analysis identified glucose (SHAP value: 2.34) and BMI (SHAP value: 1.87) as primary predictors, demonstrating strong clinical concordance. Feature interaction analysis revealed synergistic effects between glucose and BMI, providing actionable insights for personalized intervention strategies. Conclusions: Despite promising results, further validation of the proposed framework is required prior to any clinical deployment. At this stage, the study should be regarded as a methodological proof-of-concept rather than a clinically generalizable evaluation. Our framework successfully bridges algorithmic performance and clinical applicability. It achieved high cross-validated performance on a publicly available Kaggle dataset, with Random Forest reaching 96.9% accuracy and 0.998 AUC. These results are dataset-specific and should not be interpreted as clinical performance. External, prospective validation in real-world cohorts is required prior to any consideration of clinical deployment, particularly for personalized risk assessment in healthcare systems.

Over the last six decades, significant progress has been made in studying poverty. Poverty research is an important issue for ensuring the sustainable development and governance of the world, especially the lower-income regions. More attention from multiple dimensions shall be paid to poverty research. However, apart from a few research publications, less research has investigated their citations using the most widely used approach for citation mapping—Scientometric analysis. Therefore, this research utilized 319 published papers on poverty research from 1964 to 20 February 2022, from the Scopus database to conduct bibliometric and social network analysis. The bibliometrix package in R and the VOSviewer program were used to perform data analysis and visualization. Theme mapping, trend themes, bibliometric coupling, and co-occurrence networks were utilized to discover potential study areas for existing and future trends. The findings reveal that poverty research has increased by 10.18% each year since 2006. Additionally, the results indicate the most influential sections of the research based on the most often mentioned subjects, papers, authors, and keywords. The findings indicate that future studies should focus on the poverty line, social policies, and living standards. The contributions of the paper may provide a reference to the understanding of poverty research through bibliometric analysis and promote poverty research in theory and practice.

This research evaluates the performance of various sentiment analysis models, including traditional machine learning approaches (Naive Bayes, KNN, CART), a deep learning model (LSTM), and the transformer-based model RoBERTa using an Amazon book reviews dataset. ROBERTa outperformed all other models, achieving an accuracy of 96.30% and an F1-score of 98.11%, underscoring its superior ability to process complex and semantically diverse textual data. Traditional models, while computationally efficient, demonstrated limitations in capturing nuanced textual relationships, and the LSTM model, although competitive, faced scalability challenges and overfitting issues. These results demonstrate how transformer-based architectures such as RoBERTa offer advantages in real-world applications, particularly in e-commerce and social media sentiment analysis. This study underscores the superior capabilities of RoBERTa for sentiment analysis, particularly in processing semantically diverse and context-rich textual data that traditional models struggle to capture. Future work will explore optimizing RoBERTa's computational efficiency and expanding its applications to multilingual and cross-domain sentiment analysis tasks.

It is very challenging to regulate emerging innovations in the rapid evolution of online education, especially during the pandemic of Covid-19. Modern online learning systems must take both educational and technological factors into account. For learning purposes, universities have attempted to facilitate their students with online learning practices and phases using emerging technologies. In this study, the online student platform for humanities and social sciences students (HUSO-OPS) was developed using the user-centric, TPACK, and V-model design. Especially, the V-model mostly focused on the field of app engineering online learning network coordination. The 5-axis framework was to enhance the TPACK model to address online learning in more detail about pedagogy, content, movement technology, and the learner profile. The development also concerns the content as well as the context of learning. This research demonstrates the outstanding use of the HUSO-OPS platform during the global pandemic of Covid-19. They explain the different factors involved and the successful interconnections needed for optimum and efficient online learning implementation. The balance of the five parts on the online learning platform in design or consultation is a plural reflection matter.

This study demonstrates the effectiveness of advanced machine learning techniques in detecting fraudulent activities within the banking industry. We evaluated the performance of various models, including LightGBM, XGBoost, CatBoost, vote classifiers, and neural networks, on a comprehensive dataset of banking transactions. The CatBoost model exhibited the highest accuracy in identifying fraudulent instances, showcasing its superior performance. The application of diverse sampling and scaling techniques significantly improved fraud detection accuracy, emphasizing their crucial role in the process. Furthermore, the incorporation of the CatBoost ensemble method substantially enhanced the efficiency of fraud identification. Our findings underscore the potential of these advanced machine-learning approaches in mitigating financial losses and ensuring secure transactions, ultimately bolstering trust and security in the banking sector. Future research directions include refining the CatBoost model’s hyper parameters, adapting to evolving fraud patterns, and integrating real-time data for enhanced responsiveness. Additionally, efforts will be made to improve the interpretability of the model’s decision-making process, providing valuable insights into its trust-building capabilities and enhancing the transparency of fraud detection methodologies.