ABSTRACT This paper introduces an innovative Light‐Weight Consensus model for Context‐Sensitive Side Chains (abbreviated as LWC2S2C) to enhance the efficiency of consensus mechanisms within blockchain networks. The proposed model adopts a multifaceted approach, taking into account many intricate parameters including the number of blocks within the sidechain, the architecture of each block, the performance metrics of individual miners, and the complicated interplay between source nodes and miner nodes. These various metrics are combined to give each miner a unique rank, which affects their chances of being chosen for the important block validation process. The consequential discoveries derived from this comprehensive analysis reveal that the LWC2S2C model eclipses its contemporary counterparts in consensus mechanisms. Through the meticulous examination of performance benchmarks, including mining delay, energy consumption, and throughput, within diverse application scenarios, namely Electronic Health Records (EHR), Internet of Medical Things (IoMT), and Enterprise Resource Planning (ERP), the LWC2S2C consistently manifests an enviable supremacy. Remarkably, it outpaces the Practical Byzantine Fault Tolerance (PBFT) by an impressive 10.5%, surpasses the Proof of Practicality and Trust (PoPT) by a striking 14.6%, and leaves the Improved Proof of Trust (IPoT) behind by an astonishing 18.4% in the context of mining delay. Furthermore, in the case of energy efficiency, the LWC2S2C model consumes 6.2% less energy compared to PBFT, 10.5% less energy compared to PoPT, and a substantial 16.8% less energy when contrasted with IPoT. These findings emphatically underscore the scalability, efficiency, and low energy footprint that the LWC2S2C model brings to the fore. The obtained results demonstrate that LWC2S2C reaches highly awarding inference tailored to the exigencies of sidechain‐based applications.

Fuzzy cognitive maps (FCMs) have been proven effective in modeling and predicting stationary time series, yet challenges persist when dealing with time-varying nonstationary time series characterized by dynamic statistical features. This article presents a robust hybrid predictive approach, which combines an improved version of empirical Fourier decomposition (IEFD) with high-order intuitionistic fuzzy cognitive maps (HIFCM), termed IEFD-HIFCM, to address these challenges in time-series forecasting, focusing on manufacturing applications. IEFD-HIFCM offers three key contributions to overcome existing limitations in the FCM-based time-series forecasting literature. First, we introduce IEFD to extract features from the original time series that later to be fed into the HIFCM, addressing the shortcomings of established methods, such as empirical wavelet transform, variational-mode decomposition, and Fourier decomposition. Second, by using HIFCM, the approach possesses an answer for uncertainty by considering the degree of hesitation between nodes in the cognitive map. Third, this article combines elastic-net with an enhanced version of the grey wolf optimizer to optimize the weights and parameters of HIFCM as a whole, rectifying the issue with earlier FCM-based predictors that optimize individual components separately. IEFD-HIFCMs performance is validated through comparisons with state-of-the-art methods using a mathematically generated nonstationary signal. Additionally, the proposed approach is tested on four real-world smart manufacturing and supply chain datasets, yielding highly accurate results. These results demonstrate the effectiveness of IEFD-HIFCM in enhancing time-series forecasting accuracy and reducing forecasting errors.

The emergence of new technologies has had a noteworthy impact on communication systems, leading to the importance of conducting research in this area due to the significant influence of social media. Marketing operators must implement the necessary infrastructure to identify and fulfill customers' expectations, considering the advantages and the increasing number of users in social networks. This paper presents a comprehensive framework for evaluating the purchase intention of electronic commerce systems, taking into account the impact of social media advertising and customer loyalty. The paper aims to enhance purchase intention in e-commerce through social media advertising and examine influential factors that improve online shopping performance, including social media advertising's effectiveness on customer purchase behavior and brand loyalty. The paper develops a theoretical framework of nine hypotheses to evaluate purchase intention in e-commerce systems, with a focus on the effect of social media advertising on brand loyalty and purchase intention. To validate the proposed model and test the research hypotheses, we make use of information obtained from an international company located in China. The results demonstrate the positive effect of pleasurable motivation on purchase intention, with a significance level of 3.776 and a path coefficient of 0.279. Moreover, an investigation of the positive effect of customer loyalty on the recommended advertisement confirms the hypotheses, with a significance of 32.815 and a path coefficient of 0.788.

Summary In today's world, many public and private services are provided virtually on the Internet. Due to the increasing dynamism and development of computer networks, intrusion detection systems, as one of the hottest topics in network security, has become an attractive area of research for researchers. The intrusion detection system tries to categorize the activity of the connections into two categories, normal and abnormal. In intrusion detection system, each connection is described based on a set of features, and decisions about whether that connection is normal or abnormal are made using those features. The act of determining the norm or abnormality of a connection is called classification. In this article, a method based on combined classification is proposed to detect zero‐day attacks. One of the most important innovations in this method is using a new version of the GRASP feature selection algorithm, which is used to diversify the base classifiers. In this method, an attempt is made to produce a subset of different features that have high accuracy; and variety to be used in the assembly stage. Experimental results showed that the method used to create feature subsets has high quality.

Abstract This article proposes a swarm intelligence ensemble deep transfer learning (named SI‐EDTL) for multiple vehicle detection in unmanned aerial vehicle (UAV) images. This method is based on Faster regional‐based convolutional neural networks (Faster R‐CNN), in which, a set of region proposals are extracted using region proposal network (RPN), and then, CNN is used to mine highly descriptive features of these windows to classify regions. We use three Faster R‐CNNs as feature extractors (InceptionV3, ResNet50, and GoogLeNet) that have already pre‐trained on ImageNet data, combined with five transfer classifiers (KNN, SVM, MLP, C4.5 Decision Tree, and Naïve Bayes). As a result, 15 different base learners are trained through deep transfer learning on a UAV dataset to classify the region proposals into multiple vehicles (car, van, truck, and bus). We combine these 15 base learners through a weighted averaging aggregation into four vehicle classes or no vehicle (background). Hyperparameters of the ensemble model are tuned using whale optimization algorithm, to achieve the best trade‐off between total accuracy, precision, and recall. The proposed SI‐EDTL model has been successfully developed using parallel processing in MATLAB R2020b. Experimental results on AU‐AIR dataset of UAV images demonstrate the superiority of the SI‐EDTL model against existing techniques.

Abstract Given the growing use of e‐learning and expansion of internet‐based infrastructure during COVID‐19 epidemic, the need for a resilient approach to e‐learning systems is deeply felt. This article introduces a combined technique utilizing adaptive neuro‐fuzzy inference system (ANFIS) and genetic algorithm (GA), named ANFIS‐GA, to evaluate e‐learning resilience. In the proposed ANFIS model, 22 features from five main factors including individual, technology, content, agility, and assessment/support factors are used as fuzzy inputs, while the e‐learning resilience is considered as a single output of the model. To select the most significant features for the evaluation of the e‐learning resilience, an evolutionary feature selection based on GA is used. The proposed ANFIS‐GA model has been successfully developed for evaluation of e‐learning resilience in virtual Iranian university. According to the obtained results, agility is the most important factor, and then, technology and assessment/support factors have the next priorities to evaluate e‐learning resilience in virtual Iranian university. Statistical analysis demonstrated that there is no significant difference between the experts' opinion and the resilience obtained via the proposed model. The proposed ANFIS‐GA model can be used in any educational institution to evaluate the improvement of resilience in e‐learning.