Credit card fraud encompasses illicit activities aimed at unlawfully obtaining confidential information to enable unauthorized individuals to engage in illegal transactions. As technology advances, fraudsters have honed their skills in evading security measures, presenting a formidable challenge in fraud detection. To address this issue, an array of algorithms and analytical techniques has emerged to identify and mitigate instances of fraud. This research aimed to identify the most appropriate supervised machine learning algorithm for credit card fraud detection. Logistic Regression, Random Forest, Support Vector Machine, and Decision Trees were implemented and compared. Due to the imbalanced nature of the dataset, the SMOTE (Synthetic Minority Oversampling Technique) technique was employed to rectify the data imbalance by oversampling the minority class. The performance of the trained models was evaluated using various metrics, including the confusion matrix, accuracy, precision, recall, f1-score, Matthews Correlation Coefficient (MCC), and Area Under the Curve (AUC). The results of the analysis revealed that Random Forests exhibited exceptional performance, achieving an impressive recall score of 84% and surpassing other algorithms. This research provides the groundwork for future investigations involving diverse deep-learning techniques applied to real-time and dynamic datasets, enabling continuous enhancements in fraud detection and prevention mechanisms.

The effects of transcutaneous electroacupuncture stimulation (TEAS) on large-scale brain function remain insufficiently characterized. This study employed a graph-theoretical approach to analyze electroencephalogram (EEG) data from 48 healthy participants in the Pilot-6 TEAS study. Participants received sham (0 pps), 2.5 pps, 10 pps, and 80 pps stimulation during baseline, stimulation, and recovery phases. Functional connectivity was assessed using coherence and the weighted phase-lag index, followed by calculation of global and nodal graph measures from thresholded weighted undirected sensor-level networks. Descriptive analysis indicated potential frequency-related differences in EEG network organization. The 2.5 pps condition exhibited the highest average degree, whereas the 80 pps condition demonstrated the highest average clustering coefficient. At 10 pps, sensor-level maps revealed a distinct frontal–central betweenness-centrality pattern. Although 48 participants provided usable EEG data for descriptive analysis, only 3 participants had complete matched graph-metric data for all four stimulation conditions, limiting repeated-measures statistical validation. After correction for multiple comparisons, no statistically significant frequency-related effects were observed, and nodal hub differences were not independently confirmed. Consequently, these patterns should be interpreted as descriptive and exploratory rather than established group-level effects. These findings indicate that graph-theoretical EEG analysis may facilitate the identification of candidate network features for future investigations of TEAS-related brain network organization.

Data mining aims to find relevant information for decision-making, forecasting, optimising, and other business or research reasons. In this study, data mining is used in the domain of decision-making. Various problems for this domain exist in the literature, including selection problems, considering unnecessary attributes, and removing irrelevant attributes in the dataset by applying different preprocessing techniques. For this purpose, a smartphone dataset is used, and different machine learning classifiers are applied. Decision Tree, Naive Bayes, SMO, bagging, and Random Forest were chosen for precision, recall, and F-measure. Results demonstrate that Random Forest obtains 57% accuracy and performs better on average than the other algorithms. The class "Saving Account" was classified with 52% accuracy as of the other attributes, but it has the fewest errors depending on the attributes. This study can be extended by applying the proposed methodology to a rich volume dataset and deep-learning techniques.

Importance of Study: Quick and accurate detection of intracranial hemorrhage (ICH) is crucial, as delays can lead to higher risks of death, disability, and more complex treatment. While non-contrast CT scans are the main tool for diagnosing ICH, fast interpretation relies on radiologists, who may not always be available in emergencies or in places with limited resources.Novelty Statement:This study introduces NeuroAlert, a lightweight convolutional neural network (CNN) that can detect different types of intracranial hemorrhage from non-contrast CT images. Unlike more complex deep learning systems, NeuroAlert maintains high accuracy while running efficiently with TensorFlow Lite.Methodology:A framework named NeuroAlert is presented using the RSNA Intracranial Hemorrhage dataset, which includes about 750,000 CT images labeled for five types of hemorrhage and an extra category for any hemorrhage. The images were preprocessed with Hounsfield Unit windowing, normalization, resizing, and data augmentation to highlight important features. NeuroAlert uses a compact CNN with four convolutional blocks, a fully connected layer, and a softmax output layer. To handle class imbalance, weighted categorical cross-entropy was usedduring training. The final model was optimized with TensorFlow Lite for faster, lightweight use.Results and Discussion:NeuroAlert achieved a macro-averaged accuracy of 93.7%, an F1-score of 0.926, and an AUC of 0.985, demonstrating strong overall discrimination across all hemorrhage categories. Class-wise analysis showed the highest performance for the any hemorrhage class, with an F1-score of 0.95 and an AUC of 0.992, while Intraparenchymal Hemorrhage (IPH) also showed strong recognition with an F1-score of 0.94 and an AUC of 0.990. After TensorFlow Lite optimization, the model size was reduced by 61%, and inference time remained below 1 second per image, indicating near real-time performance with efficient computational requirements.Conclusion:NeuroAlert demonstrates that a carefully designed lightweight CNN can provide reliable multi-class intracranial hemorrhage detection while maintaining deployment-oriented efficiency. The proposed framework offers a promising solution for AI-assistedCT interpretation in time-sensitive and resource-limited healthcare settings.

Background:The study investigates brain tumor diagnosis using MRI, a fundamental task in neuro-oncology, as accurate tumor type identification determines diagnostic outcomes and treatment strategies. Deep learning techniques have achieved substantial success with Convolutional Neural Networks (CNNs), which primarily extract spatial features while often overlooking other essential information, such as frequency and multi-scale information.Method:The study proposes a hybrid multi-domain framework that combines spatial features from a Pretrained ResNet-50 model with frequency-domain data obtained via the Discrete Cosine Transform (DCT) and multi-scale data from the Discrete Wavelet Transform (DWT). A channel attention mechanism fuses the extracted featuresby dynamically selecting the most discriminative ones. The model uses 5-fold stratified cross-validationto assess its performance on the TCIA MU-Glioma-Post dataset.Results:The proposed hybrid model reached an overall classification accuracy of 97%, with weighted precision, recall, and F1-score all at about 0.97. It also showed strong performance across tumor types, with AUC-ROC values close to 0.99. Compared to baseline models, this framework improved accuracy by 3-5%, supporting the value of integrating features from multiple domains.Discussion:The combination of spatial, frequency, and multi-scale representations emerges as a superior approach for MRI data classification, as it captures complementary information from the data. The attention mechanism enhances the model's ability to adaptively weight each featureduring processing. The findings suggest that medical image analysis benefits from multi-domain feature fusion, and the developed framework demonstrates high-precision classification of brain tumors