Topic modeling (TM) is an unsupervised technique used to recognize hidden or abstract topics in large corpora, extracting meaningful patterns of words (semantics). This paper explores TM within data mining (DM), focusing on challenges and advancements in extracting insights from datasets, especially from social media platforms (SMPs). Traditional techniques like latent Dirichlet allocation (LDA), alongside newer methodologies such as bidirectional encoder representations from transformers (BERT), generative pre-trained transformers (GPT), and extra long-term memory networks (XLNet) are examined. This paper highlights the limitations of LDA, prompting the adoption of embedding-based models like BERT and GPT, rooted in transformer architecture, offering enhanced context-awareness and semantic understanding. The paper emphasizes leveraging pre-trained transformer-based language models to generate document embedding, refining TM and improving accuracy. Notably, integrating BERT with XLNet summaries emerges as a promising approach. By synthesizing insights, the paper aims to inform researchers on optimizing TM techniques, potentially shifting how insights are extracted from textual data.

Objectives: The primary objective of the study is to propose and evaluate a novel solution to address network congestion in VANET technology by introducing Virtual Safety Regions (VSR). This involves implementing the K-Neighborhood Network (KNN) algorithm to create virtual regions around vehicles, allowing them to connect directly with others in the same VSR without broadcasting messages across the entire network. The ultimate aim is to significantly reduce congestion while maintaining essential data transmission for safety and traffic control within VANETs. Methods: The study\'s methodology revolves around implementing the K-Neighborhood Network (KNN) algorithm to establish Virtual Safety Regions (VSR) around vehicles in VANETSs. This innovative approach enables vehicles to communicate exclusively with others located within the same VSR, eliminating the need for broadcasting messages to the entire network. The study extensively examines the VSR method, delving into algorithmic intricacies and assessing its performance through simulation-based experiments. These methods are aimed at comprehensively evaluating the effectiveness of VSR in mitigating VANET congestion and improving network performance. Findings: The study\'s findings highlight the success of the proposed Virtual Safety Region (VSR) method in addressing VANET congestion. Through simulation-based experiments, it was demonstrated that VSR significantly reduces network congestion, leading to marked improvements in network performance and reduced communication latency compared to existing methods. Moreover, the study identified various potential applications for VSR within the transportation sector, including accident prevention, traffic management, and emergency response. It also noted that VSR could be integrated with other technologies like 5G and autonomous vehicles to enhance the efficiency and safety of transportation networks. In summary, the findings underscore the potential of the VSR methodology to effectively mitigate VANET congestion and enhance the overall dependability of VANET technology, with far-reaching implications for the future of transportation systems. Novelty: The novelty of this study lies in its holistic approach to mitigating VANET congestion through the innovative VSR concept, coupled with a thorough exploration of its algorithmic aspects, empirical validation, and recognition of its broader applications and adaptability to emerging technologies. Keywords: Congestion Control, VANET, BSM, VSR, KNN, ITS

The continuous demand for reliable and privacy‐sensitive face recognition and detection has led to improvements in machine learning and federated learning (FL) frameworks. While the centralized framework of existing solutions is beneficial for data collection, it presents significant privacy concerns and challenges due to the diversity of data types. To overcome these challenges, this study presents a new, scientifically grounded hybrid approach that combines ensemble deep learning with FL architecture. The authors achieve this by utilizing various deep learning models, such as AlexNet and vision transformers (ViTs), in conjunction with residual networks (ResNets), enhancing the overall efficiency of data prediction and enhancing the versatility of non‐independent and identically distributed (Non‐IID) datasets. The last component, FL, is the need to store personal user data in centralized repositories. Major developments include the application of secure aggregation techniques and differential privacy to avoid leakage of data, minimizing overhead communication, and designing strategic learning for different clients’ data distribution. Real‐world testing on a distributed dataset with 50,000 facial images shows that the proposed framework can achieve a recognition accuracy of up to 95.6%, which is better than the best single‐model FL systems and other similar benchmark methods. Due to the framework’s ability to process large volumes of data from noisy and imbalanced data sets, surveillance, healthcare, and the internet of things are viable application areas. This work highlights the possibility of developing such cascaded solutions using ensemble deep learning embedded in a FL architecture.

Proliferating the internet of things (IoT) across all industry fields offers numerous possibilities for invention. It also multiplied the issues of ensuring uniform interoperability among a wide range of devices and platforms. The focus of this paper is to propose an approach for enhancing the security of IoT networks. This study investigated the possible efficacy of employing a federal public key infrastructure (PKI) structure system to serve IoT-based ecosystems. To achieve this goal, we have developed an elaborate experimental framework incorporating different trust models, security protocols, privacy enhancements, and performance metrics that demonstrate the practical benefits of this kind of federated system. One of the contributions of this study is an experimental model that mimics a real IoT ecosystem. It entails many IoT devices, installing vital PKI elements, and the development of safe information transmission channels. Measurements such as latency pointed out the feasibility of various IoT concepts, including such short response time as 2.8 ms for vehicular IoT (V2X). Measures for interoperability ranged with V2X having a 96.4% success, indicating the strength of the standards within that segment. This study reveals the benefits of a federated PKI management system for solving issues of IoT interconnectedness.

Cyber-physical systems (CPS) are significant to main organizations like Smart Grids and water conduct and are gradually helpless to an extensive range of developing threats. Identifying threats to CPS is of greatest significance, owing to their progressive frequent usage in numerous critical assets. Traditional safety devices like firewalls and encryption are frequently insufficient for CPS designs; the execution of Intrusion Detection Systems (IDSs) personalized for CPS is a crucial plan for safeguarding them. Artificial intelligence (AI) techniques have shown abundant probability in numerous areas of network security, mainly in network traffic observation and in the recognition of unauthorized access, misuse, or denial of network resources. IDS in CPSs and other fields namely the Internet of Things, is regularly considered through deep learning (DL) and machine learning (ML). This manuscript offers the design of an Advanced Threat Detection utilizing the Lemurs Optimization Algorithm with Deep Learning (ATD-LOADL) methodology in the CPS platform. The primary of the ATD-LOADL methodology is to focus on the recognition and classification of cyber threats in CPS. In the preliminary phase, the pre-processing of the CPS data takes place using a min-max scaler. To select an optimum set of features, the ATD-LOADL technique uses LOA as a feature selection approach. For threat detection, the ATD-LOADL algorithm uses a multi-head attention-based long short-term memory (MHA-LSTM) classifier. At last, the detection results of the MHA-LSTM method are boosted by the use of the shuffled frog leap algorithm (SFLA). The experimentation outcomes of the ATD-LOADL approach can be widely investigated on a benchmark CPS dataset. An experimentation outcome stated the enhanced threat detection results of the ATD-LOADL technique over other existing approaches