Al-Noor Journal for Information Technology and Cybersecurity

This research examines the efficiency of de-identification techniques in enhancing privacy protections for sensitive data using Long Short-Term Memory (LSTM) models. Following a structured five-step methodology such as Dataset Collection, Data Preparation, Feature Extraction, Classification, and Performance Evaluation. The study evaluates LSTM’s performance of dataset based on Resume, Construction, and Medical domains. The primary goal is to examine the ability of de-identification methods to hide certain information based on classification accuracy. Results indicate that LSTM achieves accuracy levels 97.14% on unmodified data, explaining its success detecting sensitive information. However, after applying de-identification using Java Programming at pre-processing phase to eliminate sensitive keyword, the accuracy drops to 78.30%.These findings highlight the effectiveness of de-identification techniques to enhance data privacy, especially in fields that require strict confidentiality.

One of the major concerns in network security that pose a big challenge to safeguarding networks is distributed denial-of-service (DDoS) attacks. Such attacks often lead to breaches of trust in online systems, cause significant losses in financial markets, and deny services to legitimate users. This study aims to propose a robust method for detecting DDOS attacks accurately. To accomplish this goal, the study investigated several machine learning algorithms in detecting such attacks utilizing the CIC-DDOS-2019 dataset, a well-known benchmark dataset characterized by its comprehensive coverage of DDOS attacks. Five machine learning algorithms have been evaluated: Random Forest (RF), Naive Bayes (NB), Logistic Regression (LR), J48 Decision Tree, and XGBoost based on their performance in detecting and discriminating between DDoS attacks and benign records. The results show high detection capability, with accuracy rates above 99% for all models except for NB. The RF, LR, J48, and XGBoost algorithms can recognize intricate DDoS assault patterns. In addition to comparing several machine learning methods for DDoS detection, this study provides insight into how these models can be helpful in real-world scenarios for improving network security.

This comprehensive review explores the evolution and current state of AI-driven data mining techniques, emphasizing their transformative impact across various sectors. We delve into key algorithms, including machine learning and deep learning methods, and their applications in fields such as healthcare, finance, and marketing. By synthesizing recent advancements and challenges, this paper aims to provide an ultimate overview of how these techniques enhance data analysis, uncover hidden patterns, and drive decision-making processes.

This review paper examines the integration of Artificial Intelligence (AI) within data mining, focusing on various algorithms, tools, and applications across different sectors. The review details the strengths and weaknesses of key algorithms such as supervised learning, unsupervised learning, and reinforcement learning. Furthermore, it discusses popular data mining tools and presents case studies highlighting the impact of AI on fields like healthcare, finance, and retail.The review concludes by identifying emerging trends, challenges, and future research directions in AI-driven data mining. The review details the strengths and weaknesses of key algorithms such as supervised learning, unsupervised learning, and reinforcement learning. Furthermore, it discusses popular data mining tools and presents case studies highlighting the impact of AI on fields like healthcare.

A substitution box is designed as a confusion component to give the modern cipher strength against differential cryptanalysis, which makes the S-Box a vital and only nonlinear part of the most modern cipher algorithm. Encryption methods that are based on chaos systems are very popular because they display a similar property to cryptography. However, most of the recently designed S-Boxes are focusing on some criteria leaving others, harder to implement, or slower to generate. Therefore, this paper proposes a dynamic design methodology to generate a chaotic S-box by utilizing the Jellyfish Search algorithm after modifying it to fit the purpose and that can be used in modern encryption algorithms.The statistical analysis results of the proposed S-Box are compared to some of the recently designed 4 × 4 S-Boxes in the literature, the comparison showed that the suggested S-Box has mostly equal, to better statistical attributes, which mark the suggested S-boxes robust cryptographically and a good fit to be used for lightweight block cipher algorithms.

The development of Informatics fields and data transfer through the internet increases daily. This development has emerged in the need to protect this data by developing cryptographic algorithms and concealment techniques to reach a higher level of protection. This work is developed using the Least Significant Bit (LSB). Focusing on changing concealment sites in the less important cells between points, and also using a suggested algorithm to encrypt grayscale image data by adopting the positivist method of image point locations. The algorithm was applied to more than one grayscale image and the data was hidden in more than one color image. The algorithm was approved and achieved good results after using metrics. The Peak Signal Noise Ratio (PSNR), The Mean Squared Error (MSE), and the Normalization Correlation (NC) standard measure the accuracy of extracted data.

The topic of security for computers is of significant importance. Over the past decade, countless cybercrimes have been executed by exploiting software flaws. This issue has led to considerable social stress, substantial losses, and higher interest in security. Vulnerabilities in applications developed in various programming languages can be identified using various methodologies and techniques. We can employ static or dynamic methods for analysis to detect vulnerabilities. Bandit is a tool for static analysis designed to identify security vulnerabilities in Python code, examining a defined range of issues. This study introduces an additional collection of vulnerabilities, specifically the top 25 CWE, to enhance the tool's detection capabilities. The approach involves analyzing Python code and constructing an Abstract Syntax Tree (AST) using the AST library in Python. By traversing the nodes of the tree and gathering information regarding the code's characteristics, potential vulnerabilities are identified based on predefined checks for each scenario. The tool's capability for predicting all the incorporated scenarios was demonstrated after the completion of the tests added to it.