D2H-AD: A Hybrid Model Utilizing Hyperdimensional Computing for Advanced Anomaly Detection

Accurate rainfall forecasting is crucial in activities such as agricultural planning, water resource management, and flood preparedness, so failure to accurately forecast rainfall might undermine these activities. This study aimed to improve the accuracy of daily rainfall prediction in Hai district, Kilimanjaro region, Tanzania, a region characterised by a bimodal rainfall pattern, by comparing rainfall prediction accuracies of Traditional Machine Learning (TML) and Deep Learning (DL) models. The 25-year daily rainfall dataset for Hai district was first pre-processed by cleaning, normalisation, and splitting into training (first 60%), validation (next 20%), and test (last 20%) sets. TML models (K-Nearest Neighbours (KNN) through its variant KNN Regressor and Support Vector Machine (SVM) through its variant Support Vector Regressor (SVR)) were developed, trained, and their performances compared with developed and trained DL models (Long Short-Term Memory (LSTM) and Gated Recurrent Units (GRU)). Each model’s test performance was evaluated using the Root Mean Squared Error (RMSE) metric. The findings revealed that the DL models, LSTM and GRU, achieved test RMSE scores of 6.35 mm and 6.49 mm, respectively, while the TML models, KNN and SVM, achieved test RMSE scores of 6.61 mm and 10.42 mm, respectively. The study demonstrated that DL models are more effective for accurate rainfall forecasting than TML models in regions with complex rainfall patterns. The insights gained will guide the development of more reliable, location-specific weather prediction systems, ultimately supporting informed decision-making in agriculture, environmental management, and flood risk reduction in Tanzania and other regions facing similar climatic challenges

BACKGROUND Patient heterogeneity is a key challenge in immunotherapy for IDH-wildtype glioblastoma (GBM), where hypermutated tumors or highly immunosuppressive tumor microenvironments (TME) act as clinical outliers affecting treatment response and trial stratification. We developed Alert-GBM, a radiogenomic AI model designed to flag such outliers based on radiomic signatures from diagnostic imaging. MATERIAL AND METHODS We used multicenter open-access radiogenomic datasets (n = 176) to train unsupervised outlier detection models. Principal Component Analysis (PCA) served as a first-pass filter to harmonize data before applying Isolation Forest (IF) modeling. Three IF models were trained using inter-dataset leave-one-out cross-validation (LOOCV) to simulate real-world performance and were benchmarked against each other. The final model was selected using stratified 10-fold cross-validation (CV). Outliers were validated through visual inspection, statistical profiling, comparison with other outlier detection methods (Local Outlier Factor, One-Class SVM, Mahalanobis distance), and correlation with biological and immune profiles from imaging and omics data. RESULTS The PCA filter identified 3 initial outliers based on MRI metadata. The final IF model, following stratified 10-fold CV, detected 6 outliers across the datasets, with anomaly scores plotted. Flagged cases included large, highly enhancing tumors, minimally enhancing tumors, and post-operative scans. Visualizations using PCA, t-SNE, and UMAP were generated to illustrate separability. False positives and negatives were assessed using immune-based scores, and the distribution of anomaly scores was analyzed with KDE plots. Cross-validation against alternative outlier models confirmed consistency of the flagged cases. CONCLUSION Alert-GBM is a proof-of-concept AI architecture for outlier detection in GBM MRI radiogenomics, capable of flagging cases with atypical immune TME profiles. Integrating such models with radiogenomic immune predictors at the time of diagnosis may support multidisciplinary decision-making and improve patient stratification for immunotherapy trials. Incorporating a human-in-the-loop framework ensures that AI outputs remain clinically interpretable and actionable.

Accurate energy demand prediction is crucial for efficient grid management and resource optimization, particularly across multiple countries with varying consumption patterns. However, real-world energy demand data often contains outliers that can distort forecasting accuracy. This study evaluates the impact of five outlier detection techniques—Z-Score, density- based spatial clustering of applications with noise (DBSCAN), isolation forest (IF), local outlier factor (LOF), and one-class support vector machine (SVM)—on the performance of three time-series forecasting models: long short-term memory (LSTM) networks, convolutional neural network (CNN) Autoencoders, and LSTM with attention mechanisms. The models are tested using energy demand data from four European countries— Germany, France, Spain, and Italy—derived from real-time consumption records. A comparative analysis based on root mean squared error (RMSE) demonstrates that incorporating outlier detection significantly enhances model robustness, reducing forecasting errors caused by anomalous data. The findings emphasize the importance of selecting appropriate outlier detection strategies to improve the accuracy and reliability of energy demand forecasting. This research provides valuable insights into the trade-offs involved in outlier removal, with implications for policy and operational practices in energy management.

Intrusion detection plays a critical role in safeguarding computer networks from malicious activities and unauthorized access. Traditional machine learning (ML) models have been widely employed for intrusion detection, but their effectiveness can be limited due to the complex and evolving nature of network attacks. An efficient Convolutional Neural Network (CNN) that utilizes one-dimensional convolutions, batch normalization, and max pooling layers is designed to capture the attack patterns in network traffic data. By leveraging the inherent feature extraction capabilities of CNNs, the proposed architecture demonstrates improved performance in identifying network intrusions compared to traditional ML models. The effectiveness of the novel CNN architecture is evaluated by conducted comprehensive experiments on two benchmark intrusion detection datasets. The designed CNN performance was compared with four other ML models such as logistic regression, k-nearest neighbors, decision trees, and random forest. The experimental results revealed that the proposed CNN architecture outperformed the traditional ML models with highest accuracy of 97.68% and 97.39% and highest F1-score of 97.68% and 97.40% for. UNSW-NB15 and CICIDS2017 datasets, respectively. The proposed CNN architecture serves as a promising alternative to traditional ML models, offering improved performance in identifying and mitigating network attacks.

Nowadays mobile positioning devices, sensor technologies, machine learning methods and cloud computing allow an efficient collection, processing and analysis of dynamic position information (trajectories), which can be used for various applications. Currently, data owners (e.g. vehicle producers or service operators) are reluctant to share data due to data privacy rules or because of the risk of sharing information with competitors, which could jeopardize the data owner's competitive advantage. In this paper, we compare various state-of-the-art anomaly detection methods like one-class support vector machine, isolation forest and bidirectional generative adversarial networks towards the detection of abnormal vehicle trajectories at intersections solving one-class classification problem with unsupervised learning algorithms. The experiments show that the selected models allow to identify anomalies with 98%-99% accuracy for the centralized approach. However, data exists mostly fragmented in form of isolated islands throughout the whole industry. Thus, we focus on a general concept of collaborative learning which provides with an increasing number of partners more accurate anomaly detection models than local models of each partner without exchanging raw data, but only synchronizing model parameters. We propose a federated learning algorithm for an one-class support vector machine model towards the detection of anomalies. The federated collaborative approach allows to construct comparable accuracy with the centralized approach, reduces the local data labeling load and keeps individual data private.

Research suggests that return and refund fraud poses a significant challenge to the fashion e-commerce, leading to financial losses and operational inefficiency. Traditional anomaly detection methods struggle with class imbalance and inconsistency across a single model approach. In this work, we propose a novel Hybrid Stacking Ensemble Framework which integrates Isolation Forest, One-Class Support Vector Machine (OCSVM), and Autoencoder with a Logistic Regression meta-classifier. To address the severe imbalance in the dataset, SMOTE was used to generate synthetic samples for minority class. Experiment demonstrated that while baseline imbalanced dataset achieved an accuracy of 76% and F1-score of 0.76, the resulted dataset with SMOTE achieved approximately 0.92 F1-score. This paper contributes a hybrid ensemble and data augmentation pipeline that can serve as a foundation for suture anomaly detection research.

Credit card fraud is a major problem in today’s financial world. It induces severe damage to financial institutions and individuals. There has been an exponential increase in the losses due to fraud in recent years. Hence, effectively detecting fraudulent behavior is of vital importance for either financial institutions or individuals. Since credit fraud events account for a small proportion of all transaction events in real life, the datasets about credit fraud are usually imbalanced. Some common classifiers, such as decision tree and naïve Bayes, are unable to detect fraud. Furthermore, in some cases, traditional strategies for dealing with an imbalanced problem, such as the synthetic minority oversampling technique (SMOTE), are not effective for the fraud detection datasets. To accurately detect fraud behavior, this study uses anomaly detection on imbalanced data, as well as Isolation Forest (IForest) with kernel principal component analysis. A one-class support vector machine (OCSVM) with AdaBoost is used as two models to detect outliers which significantly improves detection accuracy and efficiency. The model achieved 96% accuracy, 100% precision, 96% recall, and 98% F 1 score, respectively. The proposed model is expected to become a helpful tool for finding credit card fraud detection, and the analysis presented in this study will provides useful insights into credit card fraud detection mechanisms.

The theoretical and practical aspects of an anomaly detection system in DNS queries, which serves as a crucial tool for ensuring the security of internet infrastructure, are examined. Anomalous DNS queries pose a serious threat to network security and stability as they can be an indicator of cyber attacks or malicious activity. One of the most common threats is the use of DNS to carry out DDoS attacks, in particular through the DNS amplification mechanism. An analysis of existing anomaly detection methods, including statistical, signature-based approaches, and machine learning methods, is conducted. The key stages of the proposed system's development are described, including data collection, preprocessing, and analysis, as well as the creation of a normal activity profile for identifying deviations. Innovative methods based on deep learning and time series analysis open new horizons in automated detection of anomalies in DNS traffic. Deep neural networks, such as recurrent neural networks (RNNs) and convolutional neural networks (CNNs), are used to detect complex patterns in large data sets, including the textual and temporal aspects of DNS queries. The primary goal of the study is to demonstrate the effectiveness of a combined approach to DNS traffic analysis, utilizing modern machine learning algorithms such as Isolation Forest, One-Class SVM, and K-means. The proposed system achieves a high level of accuracy (92%) and completeness (90%) in anomaly detection, as confirmed by testing results on the CAIDA Passive DNS Dataset. A description of the modular architecture of the system is presented, which allows for scalability in large networks with high traffic levels. The proposed approach is flexible and adaptive, enabling integration with existing network security and incident response tools.

Diabetic retinopathy (DR) is one of the leading causes of vision loss in adults and is one of the detrimental side effects of the mass prevalence of Diabetes Mellitus (DM). It is crucial to have an efficient screening method for early diagnosis of DR to prevent vision loss. This paper compares and analyzes the various Machine Learning (ML) techniques, from traditional ML to advanced Deep Learning models. We compared and analyzed the efficacy of Convolutional Neural Networks (CNNs), Capsule Networks (CapsNet), K-Nearest Neighbor (KNN), Support Vector Machine (SVM), decision trees, and Random Forests. This paper also considers determining factors in the evaluation, including contrast enhancements, noise reduction, grayscaling, etc We analyze recent research studies and compare methodologies and metrics, including accuracy, precision, sensitivity, and specificity. The findings highlight the advanced performance of Deep Learning (DL) models, with CapsNet achieving a remarkable accuracy of up to 97.98% and a high precision rate, outperforming other traditional ML methods. The Contrast Limited Adaptive Histogram Equalization (CLAHE) preprocessing technique substantially enhanced the model’s efficiency. Each ML method’s computational requirements are also considered. While most advanced deep learning methods performed better according to the metrics, they are more computationally complex, requiring more resources and data input. We also discussed how datasets like MESSIDOR could be more straightforward and contribute to highly evaluated performance and that there is a lack of consistency regarding benchmark datasets across papers in the field. Using the DL models facilitates accurate early detection for DR screening, can potentially reduce vision loss risks, and improves accessibility and cost-efficiency of eye screening. Further research is recommended to extend our findings by building models with public datasets, experimenting with ensembles of DL and traditional ML models, and considering testing high-performing models like CapsNet.

The Turkish Music Emotion Dataset in the CL database is a dataset for music emotion classification, which contains many music samples and their corresponding emotion labels. This paper presents a comparative analysis of the performance of the traditional random forest machine learning model and the XGBoost model on the given dataset. The findings indicate that the traditional random forest machine learning model outperforms the XGBoost model in terms of accuracy, accuracy, and recall rate. The traditional random forest machine learning model achieves an accuracy of 80.8%, whereas the XGBoost model achieves an accuracy of 75%. The recall rate of the traditional random forest machine learning model is 80.8%, whereas the recall rate of the XGBoost model is 77.2%. The traditional random forest machine learning model achieved an F1 score of 80.5%, whereas the XGBoost model attained an F1 score of 75.3%. The comprehensive evaluation of the XGBoost model outperforms the traditional random forest machine learning model. Generally, the traditional random forest machine learning model demonstrates strong performance in the domain of music emotion classification. It exhibits favorable qualities such as robustness and interpretability, and possesses the ability to effectively handle noise and missing data. The XGBoost model, on the other hand, can be trained and predicted quickly, with high accuracy and generalization ability. Therefore, in practical application, it is necessary to select the appropriate model according to the specific situation, and optimize and adjust it to obtain the best classification effect.

Background Artificial intelligence (AI) has revolutionized diagnostic radiology by improving accuracy, efficiency, and clinical decision-making. While numerous systematic reviews and meta-analyses have evaluated the performance of AI-based techniques, particularly traditional machine learning (ML) and deep learning (DL) models, a comprehensive synthesis of this evidence is lacking. An umbrella review is necessary to integrate and compare findings across various imaging applications, offering the highest level of evidence to inform clinical practice and policy. Objective This umbrella review aims to compare the diagnostic performance of traditional machine learning and deep learning methods in radiology by synthesizing evidence from existing systematic reviews and meta-analyses. Methods A systematic literature search was conducted in PubMed, Scopus, Web of Science, and Cochrane Library for systematic reviews and meta-analyses published between 2018 and 2024. Only peer-reviewed reviews evaluating diagnostic applications of AI in radiology were included. Methodological quality was assessed using the AMSTAR 2 tool, and risk of bias was evaluated using ROBIS. Results Seven systematic reviews and meta-analyses were included. Across multiple imaging domains—such as breast cancer screening, spinal stenosis, and tumor grading—deep learning consistently outperformed traditional ML in diagnostic accuracy and sensitivity. Evidence quality was rated moderate to high, though variability in reporting and a lack of external validation were noted. Explainability features in DL models were underutilized and poorly evaluated. Conclusion This umbrella review confirms the superior diagnostic performance of deep learning methods over traditional ML in radiology. Future research should prioritize model transparency, real-world validation, and standardization in clinical implementation.

A vibration-based 1DCNN-BiLSTM model for structural state recognition of RC beams

One-class Classification-Based Machine Learning Model for Estimating the Probability of Wildfire Risk

The high likelihood of noise and the lack of labels in Internet of Things (IoT) time-series sensor readings make the detection of abnormalities an important subject of study. When there is a high correlation between sensor data points, conventional data mining and machine learning algorithms are unable to detect abnormalities. Furthermore, the amount and speed with which IoT sensors report data is one of the reasons why traditional machine and statistical learning algorithms are unable to detect anomalies. In recent years, Artificial Neural Networks (ANN) have been widely used for developing Deep Learning models for learning and classifying unlabeled data for detecting anomalies with high accuracy. In this connection, this paper presents a study on the performance of two well-known ANN models, such as Generative Adversarial Network (GAN) and Variational Auto Encoder (VAE), and a classification model, One Class Support Vector Machine (OCSVM). To evaluate performance, simulations were conducted using a well-known network and sensor datasets. The outcomes show which technique outperformed the others in terms of detection accuracy and training duration.

Fuzzy anomaly scores for Isolation Forest