Iraqi Journal for Computers and Informatics

Data hiding is becoming increasingly important due to the growing threat to the privacy and security of data from intruders and hackers. This situation is accompanied by the advancement in artificial intelligence applications designed to reveal hidden data, making it difficult to choose the most appropriate hiding approach from those presented in literature. The benchmarking method serves as an important roadmap for making decisions. We propose a distinct plain benchmarking method called maximum error insertion (MEI) benchmarking. This approach intends to hide data using maximum error insertion. The MEI refers to the maximum amount of distortion that can be added to host data (such as image or audio) while still ensuring the successful retrieval of hidden data. The maximum error that can be generated by each hiding algorithm is intentionally inserted to the media file, thus giving us maximum error, maximum capacity, and maximum sensitivity to signal processing attacks. Investigation of the two hiding algorithms demonstrates their applicability and precision, and their implementation significantly enhances the reliability of results during the benchmarking stage.

Deep packet inspection is a network security solution that identifies and flags anomalous network traffic patterns in a local network environment. Traditional signature-based techniques for intrusion detection are limited in identifying different attacks or completely new kinds, which makes them unsuitable in some situations. In addition, most previous methods for anomaly detection have low detection rate and high false alarm. In this study, a deep packet inspection model based on support vector machine (SVM) for anomaly detection in local area networks was proposed. The proposed method combined the SelectKBest method and SVM for the categorization of anomaly in a local network environment. Results showed that the proposed method outperformed other related machine learning methods with accuracy, precision, recall, and F1-score of 94.81%, 94.03%, 94.13%, and 94.0799%, respectively. The accuracy result shows that most network traffic can be correctly identified by the SVM using the SelectKBest approach, with minimal false positives or negatives.

Boosting algorithmsarea set of machine learning techniques that arepredicated on the notion that a weak learner's acquisition of multiple basic classifiers might yield results that are superior to those of any one simple classifier used alone. A comprehensive evaluation of regularly used boosting techniques against highly investigated diseases is lacking, despite the fact that boosting approaches have been used for disease prediction in many studies. Thus, the purpose of this workis to highlight themain algorithmsand strategies in the boosting learning. The results of this work will help academics identify a more appropriate boosting approach to predict disease, as well as better understand current patterns and hotspots in diseases prediction models that use boosting learning. The results showed that adaboost algorithm outperformed other algorithms in terms of accuracy, achieving above 90%. This review also demonstrates how combining two boosting methods can increase the basic classifier's accuracy. By using AdaBoost and LightGBM, the accuracy reached99.75%. XGBoost and Gradient Boosting techniques were employed more frequently in researchesthan other boosting algorithms.

International trade is considered the central link in the complex system of contemporary international economic relations. It links all countries of the world in a unified economic system whose goal is to address economic problems at the international level through developing productive capacity, expanding employment opportunities, and enhancing the flow of production factors between countries toachieveeconomic growth. Our study aimed to clarify the considerableimpact of some macroeconomic variables (exchange rate, gross domestic product, public spending, foreign direct investment) on foreign trade in Iraqand to determine the degree of relative importance of the macroeconomic variables affecting foreign trade in Iraq. The study followed the descriptive analytical approach by collecting data related to the study for the period from 2003 to 2020 and then analyzing the data obtained through artificial neural networks.

Early diagnosis of kidney as well as pre-kidney disease is crucial for patients because it allows them to take control of their condition and could potentially avoid or delay more significant consequences that could lower their quality of life. The chance of developing a major disease might be decreased with its assistance. Almost every part of the body could be impacted by chronic kidney disease (CKD). Fluid retention in the lungs, high blood pressure, and swelling of the legs and arms are all potential side effects. This study proposes a model that makes use of machine learning (ML) algorithms for diagnosing kidney disease. The preprocessing dataset, which contains missing values and is preprocessed with the use of mean, delete, and median approaches before data scaling, is the foundation of the suggested model. To achieve the highest classification accuracy, the preprocessing stage receives the results of missing values. Support Vector Machine (SVM) and K-Nearest Neighbor (KNN) are the two classification algorithms used to classify whether kidney disease is present or absent. Classify the dataset into testing and training (40% and 60%, respectively).The accuracy, F1-score, recall, and precision have been utilized for evaluating the suggested model. The kidney disease data-set has been used to test the outcomes of the suggested model. Without preprocessing any missing values in the dataset, the algorithms SVM and K-NN obtained maximum accuracy (95% and %89). Through deleting missing values from the dataset, the algorithms SVM and K-NN obtained maximum accuracy (%96 and %93). K-NN and SVM algorithms reached a maximum accuracy of %98 when using a mean technique; when using a median method, such algorithms attained an accuracy ranging from %95 to %98.

Accurate traffic prediction poses great difficulties because of the continuously increasing scale and diversity of 5G network traffic, which is driven by user demands. Moreover, certain characteristics of 5G traffic are constantly changing; thus, simulations using traditional models often lead to incorrect estimations or inefficient utilization of available resources. Consequently, we propose a hybrid machine learning model that integrates support vector machine (SVM) and decision tree algorithms to enhance efficiency of 5G traffic prediction. The structure of the hybrid model dynamically adjusts by adding or removing hidden layers and units within the network to improve prediction performance. The efficacy of the proposed model is evaluated using metrics like mean squared error, mean absolute error, and root mean squared error (RMSE). Findings show that the hybrid model consistently achieves lower error rates than SVM alone. Further performance enhancement of the hybrid model in predicting 5G traffic is also supported by comparisons of R-squared values against signal-to-noise ratios. These outcomes show the potential of the proposed method to improve traffic prediction accuracy in 5G networks, serving as a powerful tool for network control.

Data collection and data preprocessing are crucial stages in web usage mining, mainly because of the unstructured, diverse, and noisy nature of log data. During data collection, log file datasets are loaded and merged. Effective and comprehensive data preprocessing plays a vital role in ensuring the efficiency and scalability of algorithms used in the pattern discovery phase of web usage mining. This work aims to address these phases by introducing two innovative approaches. The first approach focuses on determining the device used for accessing the web, distinguishing between computers and mobile devices. The second approach aims to determine user sessions and complete paths by utilizing the referrer URL. The entire preprocessing pipeline has been implemented using the C# programming language, and the source code is available on GitHub at the following link: https://github.com/Mohammed91/Web-Usage-Mining.

This study explores the remarkable advancements in self-driving vehicles achieved through the application of computer vision and machine learning techniques. We examine various algorithms designed for critical functions, such as object detection, image segmentation, behavior prediction, and adaptive learning, which are all integral components of autonomous driving systems. Our research highlights key performance metrics, emphasizing accuracy, efficiency, and safety. Simulated environments and real-world testing are essential for validating the effectiveness of these methodologies.
Our findings underscore the transformative potential of self-driving technology in enhancing transportation safety and its far-reaching effect on numerous industries. Notably, self-driving cars demonstrate the ability to reduce traffic accidents and improve traffic flow, which can lead to substantial economic and social benefits. Moreover, we discuss future research avenues, including the enhancement of system robustness and safety measures, the improvement of human–AI interaction, and the utilization of edge computing and edge AI. We also address the ethical and regulatory challenges associated with the widespread adoption of autonomous vehicles.
Our comprehensive analysis indicates that self-driving technology is poised to revolutionize the transportation sector, offering safer, more efficient, and more accessible mobility solutions. As technology continues to evolve, ongoing research and development will be crucial in overcoming current limitations and realizing the full potential of autonomous driving systems.

Accurate and rapid classification of diabetic retinopathy is critically significant in order to prevent vision loss. The advent of artificial intelligence introduces novel and potent methodologies for enhancing the classification of diabetic retinopathy as derived from medical imaging. Due to the large size of the model make it unsuitable in real world. This research paper is dedicated to the classification of diabetic retinopathy utilizing constrained resources while achieving elevated accuracy levels. We implemented weighted clustering technique within deep convolutional neural networks and transfer learning architectures: VGG 19, DenseNet 121, and EfficientNet B6. To mitigate the challenge posed by considerable model sizes without sacrificing accuracy, the best fit results were observed with EfficientNet-B6, where applying weighted clustering reduced the model size by a factor of 12 while maintaining high accuracy results of 92% for the APTOS-2019 data. This underscores the efficacy of employing lightweight techniques to enhance the practicality of extensive models for the early diagnosis of diabetic retinopathy.