Al-Iraqia Journal for Scientific Engineering Research

Distributed Denial of Service (DDoS) attacksare among the most dangerous types of attacks. These kinds of attacks bring targeted servers down and make their services unavailable to legal users. The first objective of this study is to identify infected Ethernet and detect various kinds of up-to-dateDDoS attacks using a dynamic threshold by implementing multiple features of entropy and the Sequential Probabilities Ratio Test approach (E-SPRT). The second is to select relevant features to improve the performance of detection by implementing a new combination of machine learning techniques,which are ANOVA, Extra Trees Classifier, Random Forest, and Correlation Matrix with Pearson Correlation approaches. Canadian Institute for Cybersecurity (CIC-DDoS2019) databases were utilisedto evaluate the implementation. ESPRT using a feature selection approach with five features achieved an accuracy of over 97% with an average False Positive Rate (FPR) close to 0 in identifying most different kindsof DDoS attacks.

Software defect prediction (SDP) is essential for improving software reliability and reducing maintenance costs. In dynamic development environments, traditional static feature selection methods often fail to adapt to evolving data patterns. This study introduces a Q-learning–based adaptive feature selection approach, integrated with a Random Forest classifier, to enhance SDP performance. The method applies a reward-driven selection process during training, dynamically identifying the most relevant features.Experiments were conducted on a real-world bug report dataset from Kaggle (136 instances, 6 features, ≈71% positive defect cases). Model performance was evaluated using accuracy, precision, recall, F1-score, and ROC–AUC. The proposed configuration achieved an accuracy of 10.71% and exhibited very low recall for minority classes, highlighting the strong impact of class imbalance. Comparative tests against conventional feature selection methods (e.g., ReliefF, mutual information) and alternative classifiers (e.g., SVM, Gradient Boosting) confirmed that the current approach underperforms state-of-the-art SDP models.Despite this, the study demonstrates a reproducible framework for integrating reinforcement learning into feature selection for SDP and identifies key improvement areas, particularly in reward function design, imbalance handling, and dataset expansion. These findings provide a foundation for developing more adaptive, imbalance-resilient defect prediction systems in future research

Theelectronic structure and some thermodynamic properties of Carvacrolmolecule were investigated in detail with calculations at MP2/6-311G(d,p) level in environments such as gas, n-octanol, acetone, ethanol, acetonitrile, DMF, water. The effects of different environments on electron-filled HOMO, HOMO-1, HOMO-2, HOMO-3, and electron-empty LUMO, LUMO+1, LUMO+2, LUMO+3, which are close to the frontier orbitals, and also on polarizability, hyperpolarizability, and thermodynamic parameters of the molecule were investigated. It was observed that highly polar solvents significantly affected the electron density and stability of Carvacrol molecule. It was found that the electronic structure and optical properties ofCarvacrolmolecule were affected by the solvent environment. Findings about NLO properties and electronic properties of Carvacrolmolecule in different environments provide important information in optoelectronic and pharmaceutical applications

Emotion recognition from EEG signals has emerged as a pivotal area of research, driven by its transformative potential in healthcare, brain-computer interfaces, and affective computing systems. However, the intrinsic complexity, non-linearity, and susceptibility to noise in EEG data present significant challenges to accurate emotional state classification. This study proposes a robust and interpretable hybrid deep learning model for EEG-based emotion recognition. The architecture integrates Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) networks, and attention mechanisms, together with advanced signal processing techniques such as Continuous Wavelet Transform (CWT) and Power Spectral Density (PSD). This integrated approach facilitates the extraction of comprehensive spatial, temporal, and spectral features from EEG signals, enhancing the model’s ability to capture intricate patterns associated with emotional states. Experimental evaluations on the SEED-IV dataset, encompassing four emotional categories—Neutral, Happy, Sad, and Fear—demonstrated the model’s exceptional performance, achieving a macro-average F1-score of 93% and an area under the ROC curve (AUC) of 0.94. These results validate the model’s effectiveness in accurately distinguishing complex emotional patterns, even under noisy conditions and inter-class ambiguities. Overall, this research advances the domain of EEG-basedemotion recognition by introducing a high-performing, interpretable framework suitable for real-world applications while laying the foundation for future developments in adaptive neurofeedback systems and emotion-aware brain-computer interfaces.

Alzheimer’s Disease is a degenerative brain disorder that progressively impairs cognitive functions, particularly memory, and poses a significant burden on individuals and healthcare systems worldwide. Timely and accurate diagnosis of AD in its early stages is crucial to enable effective interventions and support patient care. Traditional diagnostic approaches often rely on either structural brain imaging or clinical evaluation, yet using one modality in isolation limits the ability to capture the complexity of the disease. This study introduces a multimodal deep learning framework designed to integrate structural Magnetic Resonance Imaging (MRI) with comprehensive clinical data for early detection of AD. The proposed system employs a three-dimensional convolutional neural network (3D-CNN) to analyze volumetric MRI scans and a multi-layer perceptron (MLP) to process structured clinical features. To enhance representational learning, the model applies an attention-based fusion strategy, including Transformer mechanisms, which enable it to focus on the most relevant modality-specific features. Furthermore, an ensemble learning approach combines the predictions of the individual modalities and the multimodal fusion branch, significantly improving the overall diagnostic performance. While the framework's proof-of-concept validation was conducted on a simulated dataset, the results demonstrate a high degree of accuracy and offer a strong basis for its application to real-world clinical data. We demonstrate that our ensemble model achieves a superior accuracy of 97.0%and an AUC of 0.985, outperforming unimodal and non-attentive multimodal baselines. The framework's explainability, highlighted by Grad-CAM and SHAP, offers valuable insights into the model's decision-making process, a critical step towards its clinical acceptance