Because of the heavy data and communication advances, the utilization of Internet of Things (IoT) gadgets has expanded dramatically. In the improvement of IoT, Wireless Sensor Network (WSN) plays out a crucial part and involves easy keen gadgets for data gathering. In any case, such savvy gadgets have requirements regarding calculation, preparing, memory, and energy assets. Alongside such requirements, the major difficulties for WSN are to accomplish dependability with the security of communicated information in a weak climate alongside pernicious nodes. This paper intends to build up an Anomalous Intrusion Detection Protocol and Intrusion Prevention Protocol for interruption evasion in IoT dependent on WSN to expand the network time frame and information reliability. The proposed framework makes dissimilar energy-efficient groups dependent on the natural characteristics of nodes. Also, in view of the (k, n) limit related Shamir mystery sharing plan, the unwavering quality also, the security of the tangible data within the Base Station and group head are accomplished. The proposed security conspires demonstrates a trivial answer to adapt to interruptions produced by malignant nodes. The trial results utilizing the network test system Network Simulator-2 show that the proposed directing convention accomplished improvement as far as network lifetime, end-to-end delay as 24%, packet delay ratio as 30%, when contrasted and the current work under unique network characteristics.

Malware classification in cloudenvironments remains a critical challenge due to theincreasing complexity and volume of cyber threats.This paper proposes CMC (Cloud-based MalwareClassification), a novel framework that enhancesmulti-class malware classification efficiency throughthe integration of feature selection, dimensionalityreduction, and imbalanced data handlingtechniques. The CMC framework aims to improveclassification accuracy and computational efficiencyby optimizing feature representation and addressingclass imbalance, which are common issues inreal-world malware datasets. To evaluate itseffectiveness, we apply the proposed model to twopublic benchmark datasets: CMD_2024 andCIC-MalMem-2022. Experimental resultsdemonstrate that CMC outperforms existingapproaches in terms of classification accuracy,F1-score, and computational efficiency, proving itspotential for real-world deployment in cloud-basedsecurity solutions. These findings highlight theimportance of intelligent data preprocessing andfeature optimization in enhancing malwareclassification on cloud platforms.

Botnet represents a persistent threat with significant implications for information systems on the Internet. In this study, we focus on the classification of DGA botnet, which is essential for identifying infected DGA botnet types to implement corresponding preventive measures. We propose a solution based on the BERT large language model and transfer learning. Experimental results on the UMUDGA and UTL_DGA22 datasets show that the proposed solution has at least 9.14

Quan trắc mực nước, đóng vai trò to lớn trong việc dự báo, cảnh báo phòng chống thiên tai, nước biển dâng, ngập lụt đô thị, thủy điện, thủy lợi, xây dựng, giao thông … phục vụ phát triển kinh tế - xã hội, an ninh quốc phòng. Trạm quan trắc mực nước thuộc mạng lưới khí tượng thủy văn quốc gia. Đên nay, công trình quan trắc mực nước cũng rất đa dạng và phong phú, quan trắc bằng thiết bị tự động phát triển nhanh chóng. Tuy nhiên, quan trắc thủ công cũng còn khá phổ biến và hiện tại hệ thống thủy chí, giản đồ vẫn sẽ còn tồn tại lâu dài. Hệ thống mạng internet hầu hết các trạm cũng đang có sẵn, công trình, tuyến đo cũng đã có. Việc sử dụng mô hình, thuật toán, cơ chế học máy của mô hình tính toán giá trị mực nước rất quan trọng. Mô hình phù hợp sẽ giảm chi phí vận hành trong thu thập số liệu, tận dụng tối đa trang thiết bị sẵn có. Mô hình tốt, có độ tin cậy cao, dễ triển khai và phù hợp với mọi công trình quan trắc mực nước hiện có trong và ngoài ngành khí tượng thủy văn. Nghiên cứu này, trình bày về mô hình cơ bản cho các trường hợp, điều kiện khác nhau. Trong mỗi điều kiện thời tiết, từng loại hình ảnh, công trình, thủy chí, giản đồ khác nhau sẽ lựa chọn các thuật toán tính toán khác nhau, nhằm cho kết quả phù hợp nhất. Tương tự như vậy, tùy thuộc tính chất hình ảnh, điều kiện khác nhau để đưa ra một số mô hình học máy cho phù hợp. Tính toán ra giá trị mực nước là khâu then chốt của kết quả của mô hình học máy. Trong kết quả sử dụng phép so sánh kết quả của một số mô hình, cơ chế học máy, thuật toán sử dụng để đưa ra nhận xét, đề xuất, kiến nghị cụ thể để áp dụng vào thực tiễn cho từng đối tượng, bối cảnh của từng trạm đo.

Botnets continue to pose a significant threat to internet-based information systems, with Domain Generation Algorithm (DGA) botnets being particularly challenging due to their dynamic and evasive behavior. Accurate classification of DGA domains is critical for identifying botnet families and deploying effective countermeasures. While deep learning models have shown promise in this area, recent studies indicate that conventional training approaches offer limited gains in classification performance, especially in multi-label scenarios involving a large number of classes. In this paper, we propose a supervised training strategy called Prioritized Selection Learning, which enhances model performance without altering the existing architecture. The method enables the model to focus on learning from more challenging or underperforming labels during an additional training phase, improving the model's ability to distinguish between difficult cases. Experimental evaluations conducted on three widely used and reputable DGA datasets demonstrate that the proposed method consistently outperforms the baseline model in terms of accuracy, recall, and F1-score. Notably, the F1-score ranges from 88.0% to 94.4% across datasets, indicating strong generalization and robustness. The results suggest that the proposed training strategy is an effective and flexible solution for improving deep learning-based DGA botnet classification and other multi-label classification tasks.

The escalating prevalence of malware necessitates a proactive and vigilant approach to its detection and mitigation. The ramifications of a successful malware attack on cloud services can be severe, underscoring the critical importance of effective malware detection mechanisms in cloud environments. To address this pressing need, we propose a comprehensive methodology for creating a novel cloud-based malware dataset, namely the CMD_2024 dataset. This dataset integrates static and dynamic attributes, providing a robust framework for malware analysis. The CMD_2024 dataset, comprising 20,850 samples meticulously labeled into various malware categories such as Virus, Trojan, Worm, Ransomware, Adware, Miner, PUA, and Downloader, is designed to facilitate the testing and evaluation of diverse analysis tools, machine learning models, deep learning models, and security systems. We enhance the dataset’s utility and effectiveness by focusing on dynamic features, particularly system calls within the cloud, in conjunction with static attributes. To address the challenges of the imbalance towards less common malware categories in the dataset, we employed the Conditional Tabular Generative Adversarial Network to generate synthetic data, significantly improving the detection capability for these rare malware samples. The application of various machine learning and deep learning classifiers, including our proposed integrated deep learning models, yielded remarkable results, achieving 99.42% accuracy in binary classification and 86.97% in multi-class classification. These outcomes demonstrate the CMD_2024 dataset’s substantial efficacy in supporting robust malware detection within cloud environments.

This paper is devoted to investigate the dynamical behaviors of malware attack on a familiar kind of complex heterogeneous networks, namely Wireless Sensor Network, and discuss an effective immunization treatment based on fractional interconnected Takagi-Sugeno (T-S) fuzzy systems. Our approach is based on the mathematical modeling to establish a controlled fractional network-based SIRS malware propagation model that better describes the attacking behavior of malicious objects on Wireless Sensor Network. After that, we point out some qualitative properties of the proposed network-based SIRS malware propagation model such as the existence of positively invariant set, backward bifurcation and asymptotic behavior. Especially, in order to study the model's stability, we evaluate an epidemiological threshold value R0, namely basic reproductive ratio, which ensures the existence of at least one endemic equilibrium P⁎ and the local asymptotic stability of malware-free equilibrium P0. As a consequence of theoretical result, the malware-free equilibrium P0 is unstable when R0>1 and hence, the rest of this paper is to address a stabilization problem for the proposed controlled network-based model and establish some sufficient conditions related to linear matrix inequalities and positive definite matrices. Finally, we illustrate the obtained theoretical results by a computational example.

Detection and clustering of commercial advertisements plays an important role in multimedia indexing also in the creation of personalized user content. In existing methodologies, the mining techniques were text, image, audio retrieval based on knowledge based environment and commercial video retrieval based on rule-based algorithms, logo-based algorithms, recognition based methods. The quality video with enhanced accuracy has been detected using the automated commercial and general program for video detection technique. The clustering technique implements the clustering process for the entire video to the frames and the required main frames are depended on the total amount of frames. The key frames are extracted from the video sequences and the duplicate key frames are eliminated from the video sequences. The video compression feature has been optimized using hybrid end-to-end compression technique to extract the features and reconstruct the video frames. The main encoding algorithm performs the optimization to produce the encoded frames and it is further compressed using the additional encoding algorithm. The performance results show that the proposed technique has the improved performances in terms of MSE, PSNR, SSIM, compression ratio and the computation time which is compared with the related techniques.

In Passive Clustered Wireless Sensor Networks (WSNs), energy is lost in a sensor node during the data transmission. In order to avoid the energy loss due to data transmission, a data prediction technique is implemented. In this paper, we present a new multi-point data prediction technique, in which the prediction algorithm is initially implemented at both member nodes and cluster heads. The algorithm is updated to cluster head by member nodes by tracking temporal correlation of data. Neuro-Fuzzy model is used as a predictor in both member nodes and cluster heads. The simulation is performed using MATLAB and the overall energy in nodes seems to increase. The mean square error (MSE) value is reduced to greater extend.

Malware in the cloud can affect many users on multiple platforms, while traditional malware typically only affects a system or a small number of users. In addition, malware in the cloud can hide in cloud services or user accounts, making it more difficult to detect and remove than traditional malware. Information security solutions installed on servers (such as anti-malware solutions) are not considered very effective as malware (especially sophisticated solutions) can bypass the detection capabilities of these solutions. Moreover, these solutions often cannot detect new and unknown malware patterns. To address this issue, machine learning (ML) methods have been used and proven effective in detecting malware in many different cases. This chapter per the authors focuses on introducing malware detection techniques in the cloud and evaluating the effectiveness of machine learning methods used, as well as proposing an effective model to support malware detection in the cloud.

Communicable disease pandemic is a severe disease outbreak all over the countries and continents. Swine Flu, HIV/AIDS, corona virus disease-19 (COVID-19), etc., are some of the global pandemics in the world. The major cause of becoming pandemic is community transmission and lack of social distancing. Recently, COVID-19 is such a largest outbreak all over the world. This disease is a communicable disease which is spreading fastly due to community transmission, where the affected people in the community affect the heathy people in the community. Government is taking precautions by imposing social distancing in the countries or state to control the impact of COVID-19. Social distancing can reduce the community transmission of COVID-19 by reducing the number of infected persons in an area. This is performed by staying at home and maintaining social distance with people. It reduces the density of people in an area by which it is difficult for the virus to spread from one person to other. In this work, the community transmission is presented using simulations. It shows how an infected person affects the healthy persons in an area. Simulations also show how social distancing can control the spread of COVID-19. The simulation is performed in GNU Octave programming platform by considering number of infected persons and number of healthy persons as parameters. Results show that using the social distancing the number of infected persons can be reduced and heathy persons can be increased. Therefore, from the analysis it is concluded that social distancing will be a better solution of prevention from community transmission.

Abstract— Nowadays, digital transformation has brought many great changes and is becoming an essential part of real life, however, it also goes along with a considerable likelihood of being targeted in cyberattack. Indeed, the more businesses embrace digital transformation or use online services, the more opportunities hackers have to expand their cyberattacks. Hence, there is an essential need on analyzing and predicting of cyberattack on network systems. For this aim, we propose to study a controlled fractional network-based SCIRS (Susceptible - Carrier - Infectious - Recovered - Susceptible) malware propagation model and its stabilization problem based on fractional interconnected Takagi-Sugeno fuzzy system. A fuzzy state feedback controller is proposed to asymptotically stabilize the unstable malware-free equilibrium of the proposed malware propagation model and then, we establish sufficient conditions in terms of linear matrix inequalities. The effectiveness of proposed approach is illustrated by a case study of SCIRS malware propagation model on heterogeneous complex network.

The DGA botnet prevention is a burning topic in cybersecurity, with two problems: detection and classification. The DGA botnet dataset plays an essential role in the research allowing researchers to evaluate their proposed solutions. This study introduces a new dataset on DGA botnets named UTL_DGA22. Our proposed dataset not only inherits previous datasets' results but also has got own advantages. First, our new dataset includes only domain records and no other raw network traffic, helping to address the DGA botnet problem. Second, we removed duplicated botnet DGA families and added new botnet families for a total of 76 DGA botnet families presented. Third, we propose a valuable set of attributes as input for classification algorithms. Our experiments using the proposed features with several machine learning algorithms have had good results. It shows that our proposed attributes are firmly suitable for the input of the DGA botnet solution. Finally, we carefully compiled the dataset and attribute description documents to make it easy for researchers to use. The UTL_DGA22 dataset can serve as a database for researchers to develop their algorithms while objectively evaluating different solutions.

Malware in the cloud can affect many users on multiple platforms, while traditional malware typically only affects a system or a small number of users. In addition, malware in the cloud can hide in cloud services or user accounts, making it more difficult to detect and remove than traditional malware. Information security solutions installed on servers (such as anti-malware solutions) are not considered very effective as malware (especially sophisticated solutions) can bypass the detection capabilities of these solutions. Moreover, these solutions often cannot detect new and unknown malware patterns. To address this issue, machine learning (ML) methods have been used and proven effective in detecting malware in many different cases. This chapter per the authors focuses on introducing malware detection techniques in the cloud and evaluating the effectiveness of machine learning methods used, as well as proposing an effective model to support malware detection in the cloud.

In this paper, we will introduce Caputo fractional LC derivative defined in the linear correlated fuzzy-valued number space RF(A) and some of its applications, specifically the feedback control problem in the space RF(A) in the sense of Caputo fractional LC derivative. Firstly, we give definition of the Riemann-Liouville-LC integral and the Caputo fractional LC derivative of a function that takes value in RF(A). Besides, some Caputo factional LC derivative properties of the sum and difference of two functions have also been proved. Moreover, the problem of dynamic systems in space RF(A) is given in both cases where A is a symmetric or non-symmetric fuzzy number. Along with that, the stability theorems of the equilibrium point are also taken into consideration. Finally, we are interested in building a state feedback control function for the fractional differential equation system in space RF(A) to ensure the equilibrium point of the system is asymptotically stable.

Abstract— Malware in general and botnets in particular are big threats to cybersecurity. They have many sophisticated methods to bypass security systems to infect computers and perform attacks, sabotage, or spying activities. Botnet detection solutions are always focused on and solved by scientists and cybersecurity specialists. The DGA botnet is a group of common botnet families that share the same mechanism of needing to connect back to the C&C server via DNS to receive commands to operate. Many studies that propose algorithms for detecting and classifying DGA botnets have been proposed and tested with high results. In this study, we approach by using the above solutions to detect malicious IP addresses and botnet malware families. First, we evaluate the efficiency of two deep learning models LA_Bin07 and LA_Mul07 on a new specialized dataset, UTL_DGA22. Next, we extended the experiment with the ISCX-Bot-2014 dataset. The results show that LA_Bin07 and LA_Mul07 models both get high accuracy on the new dataset, with 0.98 and 0.86 correspondingly. Experimenting on the reality dataset also gives positive results, helping network administrators to localize malicious IP addresses for deeper investigation. The proposed solution is effective enough to be applied as a module in cybersecurity solutions such as firewalls, intrusion detection, and prevention systems or unified thread management - UTM.

For understanding the influence of malware attacking on complex heterogeneous networks, this work studies a fractional network-based SIRS epidemic model with fuzzy transmission and saturated treatment function. Firstly, we apply the next-generation method to obtain the basic reproductive ratio R0, that is an important threshold value in the investigation of asymptotic behavior of the proposed epidemic model. The obtained theoretical results indicates that the value R0 significantly depends on the topology structure of the underlying network and the malware load. In addition, we give a threshold value ensures the clean of malware programs on the network. At last, the sensitivity analysis of the threshold value R0 and some graphical simulations are presented to illustrate for the theoretical results. similar to R0 > R0 that not only determines the existence of endemic equilibrium E* but also

— Malware in general and botnets in particular are big threats to cybersecurity. They have many sophisticated methods to bypass security systems to infect computers and perform attacks, sabotage, or spying activities. Botnet detection solutions are always focused on and solved by scientists and cybersecurity specialists. The DGA botnet is a group of common botnet families that share the same mechanism of needing to connect back to the C&C server via DNS to receive commands to operate. Many studies that propose algorithms for detecting and classifying DGA botnets have been proposed and tested with high results. In this study, we approach by using the above solutions to detect malicious IP addresses and botnet malware families. First, we evaluate the efficiency of two deep learning models LA_Bin07 and LA_Mul07 on a new specialized dataset, UTL_DGA22. Next, we extended the experiment with the ISCX-Bot-2014 dataset. The results show that LA_Bin07 and LA_Mul07 models both get high accuracy on the new dataset, with 0.98 and 0.86 correspondingly. Experimenting on the reality dataset also gives positive results, helping network administrators to localize malicious IP addresses for deeper investigation. The proposed solution is effective enough to be applied as a module in cybersecurity solutions such as firewalls, intrusion detection, and prevention systems or unified thread management - UTM.