Human Activity Recognition (HAR) stands as a crucial technology, with applications ranging from healthcare monitoring to sports analytics. However, the traditional approach to HAR is often time-consuming and susceptible to human errors due to the high complexities involved in processing diverse sensor data. Recognizing the imperative for efficiency and accuracy in HAR systems, we propose the development of an Automated Decision-maker (ADM) system. This system serves to automate HAR pipelines, addressing the challenges posed by the huge sensor data. By harnessing the power of automation, ADM significantly streamlines the HAR process, reducing the time required for hyperparameter tuning and minimizing the risk of human errors. The results obtained from our proposed ADM system demonstrate notable improvements in HAR performance, showcasing achieved accuracy of 96.436% for UCI-HAR & 99.783% for PAMAP2 datasets. Moreover, ADM can be described as an innovative approach that contributes to the optimization of HAR systems while also establishing a foundation for building robust and reliable systems in complex environments.

The boundaries and regions between individual classes in biomedical image classification are hazy and overlapping. These overlapping features make predicting the correct classification result for biomedical imaging data a difficult diagnostic task. Thus, in precise classification, it is frequently necessary to obtain all necessary information before making a decision. This article presents a novel deep-layered design architecture based on Neuro-Fuzzy-Rough intuition to predict hemorrhages using fractured bone images and head CT scans. To deal with data uncertainty, the proposed architecture design employs a parallel pipeline with rough-fuzzy layers. In this case, the rough-fuzzy function functions as a membership function, incorporating the ability to process rough-fuzzy uncertainty information. It not only improves the deep model's overall learning process, but it also reduces feature dimensions. The proposed architecture design improves the model's learning and self-adaptation capabilities. In experiments, the proposed model performed well, with training and testing accuracies of 96.77% and 94.52%, respectively, in detecting hemorrhages using fractured head images. The comparative analysis shows that the model outperforms existing models by an average of 2.6 $\pm$ 0.90% on various performance metrics.

ABSTRACT The application of artificial intelligence (AI) in healthcare has been witnessing an increasing interest. Particularly, federated learning (FL) has become favourable due to its potential for enhancing model quality whilst maintaining data privacy and security. However, the effectiveness of present FL methodologies could underperform under non‐IID conditions, characterised by divergent data distributions across clients. The globally constructed FL model may suffer potent issues by allowing the least‐performing models to equal participation. Thus, we propose a new accuracy‐based FL approach (FedAcc) which only takes into account the clients' validation accuracy to consider their participation during global aggregation, also called Smart Healthcare Amplified (SHA). However, with limited supervised data it is challenging to increase the model performance thus concept of transfer learning (TL) is used. TL enables the global model to integrate knowledge from precomputed systems, resulting in an efficient model. However, the complexity of the global system is amplified by these TL models, leading to challenges related to vanishing gradients, particularly when dealing with a substantial number of layers. To mitigate this, we present a Transfer Learning Domain Adaptation Model (TLDAM). TLDAM employs a two‐layered sequentially trained TL model, which contains approximately 50% fewer layers compared to traditional TL models. TLDAM is trained on multiple datasets such as MNIST and CIFAR10, to enhance its knowledge and make it domain‐adaptive. Moreover, experimental results conducted on the UCI‐HAR dataset reveal the supremacy of our proposed framework with an accuracy of 94.2990%, F‐score of 94.2820%, precision of 94.3058%, and recall of 94.2993% over traditional FL techniques and state‐of‐the‐art techniques.

Recently the field of sentiment analysis has gained a lot of attraction in literature. The idea that a machine can dynamically spot the text’s sentiments is fascinating. In this paper, we propose a method to classify the textual sentiments in Twitter feeds. In particular, we focus on analyzing the tweets of products as either positive or negative. The proposed technique utilizes a deep learning schema to learn and predict the sentiment by extracting features directly from the text. Specifically, we use Convolutional Neural Networks with different convolutional layers. Further, we experiment with LSTMs and try an ensemble of multiple models to get the best results. We employ an n-gram-based word embeddings approach to get the machine-level word representations. Testing of the method is conducted on real-world datasets. We have discovered that the ensemble technique yields the best results after conducting experiments on a huge corpus of more than one million tweets. To be specific, we get an accuracy of 84.95%. The proposed method is also compared with several existing methods. An extensive numerical investigation has revealed the superiority of the proposed work in actual deployment scenarios.

Internet users perceive a multilingual web but are unfamiliar with it due to communication in their regional language called Cross-Lingual Information Retrieval (CLIR). In CLIR, a translation technique is used to translate the user queries into the target document’s language. Conventional translation techniques are based on either a manual dictionary or a parallel corpus, whereas the trending Statistical Machine Translation (SMT) and Neural Machine Translation (NMT) techniques are trained on a parallel corpus. NMT is not so mature for Hindi-English translation, according to the literature, and SMT performs better than the NMT. SMT provides a static translation due to the limited vocabularies in the available parallel corpus. It may not provide the translations for missing or unseen words, whereas the web provides a dynamic interface where multiple users are updating information at the same time. The web may provide the translations for missing or unseen words, and therefore the web is effectively used for technically developed languages like English, German, Spanish, Russian, and Chinese. In this article, different web resources such as Wikipedia, Hindi WordNet and Indo WordNet, ConceptNet, and online dictionary based translation techniques are proposed and applied to Hindi-English CLIR. Wikipedia-based translation approach incorporates three modules—exactly matched, partially matched, and disambiguation—to address the issues of wrong inter-wiki links, partially matched terms, and ambiguous articles. Hindi WordNet and Indo WorNet attribute “English synset” and ConceptNet attributes “Related term” & “Synonymy” are used for obtaining translations. Further, WordNet path similarity is used to disambiguate translations. Various online dictionaries are available that return multiple relevant and irrelevant translations. The proposed approaches are compared to the SMT where the Wikipedia-based approach achieves approximately similar mean average precision to SMT.

Agriculture is an important sector that plays an essential role in the economic development of a country. Each year farmers face numerous challenges in producing good quality crops. One of the major reasons behind the failure of the harvest is the use of unscientific agricultural practices. Moreover, every year enormous crop loss is encountered either by pests, specific diseases, or natural disasters. It raises a strong concern to employ sustainable advanced technologies to address agriculture-related issues. In this article, a sustainable real-time crop disease detection and prevention system, called CROPCARE, is proposed. The system integrates mobile vision, Internet of Things (IoT), and Google Cloud services for sustainable growth of crops. The primary function of the proposed intelligent system is to detect crop diseases through the CROPCARE—mobile application. It uses the superresolution convolution network (SRCNN) and the pretrained model MobileNet-V2 to generate a decision model trained over various diseases. To maintain sustainability, the mobile app is integrated with IoT sensors and Google Cloud services. The proposed system also provides recommendations that help farmers know about current soil conditions, weather conditions, disease prevention methods, etc. It supports both Hindi and English dictionaries for the convenience of the farmers. The proposed approach is validated by using the PlantVillage data set. The obtained results confirm the performance strength of the proposed system.

In smart cities, managing the energy in the Industrial Internet of Things (IIoT) IoT networks is a challenging issue. In a cluster-based IIoT, cluster heads (CHs) near the sink quickly deplete their energy as they are responsible for relaying a large amount of data. It results in isolating the sink from the network and causes an energy hole or hot spot problem. Many solutions exist for this problem, including sink mobility, unequal clustering, use of gateways, etc. Each of these solutions has its own merits and demerits. It is well known that combining clustering and energy-aware routing can significantly extend the IIoT’s lifetime. In this work, we propose an advanced energy management and harvesting system for enhancement of network lifetime of IIoT in smart cities by addressing the hot spot problems. The system is based on energy-aware clustering and routing algorithms that employ energy-harvesting nodes deployed along with the normal sensor nodes. We simulate the proposed system extensively for various cases of network topology. A couple of existing algorithms are analyzed and compared with simulated results. It is shown that the network lifetime is increased significantly compared to the existing ones.

Fingerprint recognition has emerged as one of the most reliable biometric authentication methods , owing to its uniqueness and permanence. However, the security and confidentiality of the user’s data are key considerations in modern biometric systems. In this study, we describe an intelligent computational technique for automatically validating fingerprints for identification and verification purposes. The feature vector is created by fusing Gabor filtering features with deep learning techniques like the faster region-based convolutional neural network (Faster R-CNN). This study uses linear and decision-based median filtering (DBMF) techniques to minimize visual impulse noise. Faster-R-CNN with DBMF was applied to the feature vectors to reduce overfitting problems while improving classification precision and reliability. For fingerprint matching, the Euclidean distance between the associated Harris-SURF feature vectors of two feature points is used to measure feature-matching similarity between two fingerprint images . Furthermore, for fine-tuned matching an iterative technique known as RANSAC (Random Sample Consensus) is used. The experimental results collected from the public-domain fingerprint databases FVC-2002 DB1 and FVC-2000 DB1 show that the proposed design is viable and performs well with an accuracy of 99.43%, MSE value of 43.321%, and an execution time of 3.102 ms which was more exact than existing models.

The rapid growth of data generated by several applications like engineering, biotechnology, energy, and others has become a crucial challenge in the high dimensional data mining. The large amounts of data, especially those with high dimensions, may contain many irrelevant, redundant, or noisy features, which may negatively affect the accuracy and efficiency of the industrial data mining process. Recently, several meta-heuristic optimization algorithms have been utilized to evolve feature selection techniques for dealing with the vast dimensionality problem. Despite optimization algorithms' ability to find the near-optimal feature subset of the search space, they still face some global optimization challenges. This paper proposes an improved version of the sooty tern optimization (ST) algorithm, namely the ST-AL method, to improve the search performance for high-dimensional industrial optimization problems. ST-AL method is developed by boosting the performance of STOA by applying four strategies. The first strategy is the use of a control randomization parameters that ensure the balance between the exploration-exploitation stages during the search process; moreover, it avoids falling into local optimums. The second strategy entails the creation of a new exploration phase based on the Ant lion (AL) algorithm. The third strategy is improving the STOA exploitation phase by modifying the main equation of position updating. Finally, the greedy selection is used to ignore the poor generated population and keeps it from diverging from the existing promising regions. To evaluate the performance of the proposed ST-AL algorithm, it has been employed as a global optimization method to discover the optimal value of ten CEC2020 benchmark functions. Also, it has been applied as a feature selection approach on 16 benchmark datasets in the UCI repository and compared with seven well-known optimization feature selection methods. The experimental results reveal the superiority of the proposed algorithm in avoiding local minima and increasing the convergence rate. The experimental result are compared with state-of-the-art algorithms, i.e., ALO, STOA, PSO, GWO, HHO, MFO, and MPA and found that the mean accuracy achieved is in range 0.94-1.00.

These days, the usage of machine-learning-enabled dynamic Internet of Medical Things (IoMT) systems with multiple technologies for digital healthcare applications has been growing progressively in practice. Machine learning plays a vital role in the IoMT system to balance the load between delay and energy. However, the traditional learning models fraud on the data in the distributed IoMT system for healthcare applications are still a critical research problem in practice. The study devises a federated learning-based blockchain-enabled task scheduling (FL-BETS) framework with different dynamic heuristics. The study considers the different healthcare applications that have both hard constraint (e.g., deadline) and resource energy consumption (e.g., soft constraint) during execution on the distributed fog and cloud nodes. The goal of FL-BETS is to identify and ensure the privacy preservation and fraud of data at various levels, such as local fog nodes and remote clouds, with minimum energy consumption and delay, and to satisfy the deadlines of healthcare workloads. The study introduces the mathematical model. In the performance evaluation, FL-BETS outperforms all existing machine learning and blockchain mechanisms in fraud analysis, data validation, energy and delay constraints for healthcare applications.

Smart vehicle parking is a collaborative effort of technology and human innovation where the efforts are to be minimized to save time and efforts. In smart cities it is one of the common challenges to introduce smart parking to increase parking efficiency and combat numerous issues like identification of free parking slot and real-time dynamic updation on traffic to save fuel and energy. In this work, a new cloud-based smart parking architecture is proposed that can help in predicting the available free parking slots in smart cities. Initially, the methodology collects the car count at any near by parking using Internet of Things (IoT) and Cloud-based approach. Later, the approach uses the Kernel Least Mean Square algorithm to make heuristic predictions about future vacancy using auto-regression. The proposed approach thus utilizes the online learning or model training. To validate the efficacy of the proposed work, the testing is done on the real-time dataset. The extensive numerical investigation is performed on parking lots of four international airports of a smart city in actual deployment scenarios. The experimentation has revealed superior performance of the method in terms of vacancy prediction.

Abstract In the respiratory analysis of human beings, lungs play an important role and can be used to find several associated diseases. One of such conditions is called pneumothorax which occurs due to leaking of air into pleural space and results in lungs collapse. The traditional methodology uses chest radiography with the manual intervention of radiologists for the detection and segmentation of the affected areas. This paper introduces a deep neural network based methodology for the automatic localization and segmentation of the proper region‐of‐interest (RoI). The proposed approach is based on transfer learning where the existing UNet model is extended and redesigned to a new architecture named PTXNet for RoI segmentation. In PTXNet, the traditional encoder is redesigned with the use of EfficientNet, SE‐ResNeXt50 and Xception convolutional neural network (CNN) architectures. Furthermore, residual blocks are introduced in the decoder phase and concatenation of the previous decoder stage feature maps in addition to standard global skip connections is performed. The PTXNet is trained on a dataset of more than 15,000 chest radiography and resulted in the mean dice coefficient of 84.89%. It is found that the proposed approach provides superior results than the UNet model with an increase in the mean dice coefficient of 18.76%.

Recommender systems are extensively used today to ease out the problem of information overload and facilitate the product selection by users in e-commerce market. Both privacy and security are two major concerns of the user in these systems. For the protection of the user’s rating, there are several existing works on the basis of encryption or randomization methodologies. This paper proposes a methodology that not only protects the privacy of ratings but also provides better accuracy. After applying fuzzification on the user ratings, random rotation and perturbation methods are used before being fed to the collaborative filtering system. In this process, similar users are grouped into clusters by which recommendation is made. By considering different cluster size on four different datasets, the proposed fuzzified k-Mode clustering method provides less MAE and RMSE value as compared to other k-Means and k-Mode clustering approach and also achieves the better privacy than randomized perturbation method by obtaining IVDM value i.e. 0.67, 0.61, 0.55 and 0.7.

In a distributed cloud context, distributed denial of service (DDoS) attacks are widespread. The services are rendered unavailable to legitimate users as a result of the overwhelming traffic, resulting in financial losses. There are possible obstacles, although several researchers have established various mitigation measures. Initially, Software-defined networking technology was revealed to protect businesses from DDoS attacks. DDoS attacks cause server outages and financial losses due to service unavailability. Meeting of service-level agreement with the customers remains a challenge. In this article, the scattered denial-of-service mitigation tree architecture (SDMTA) is used to propose a novel DDoS mitigation strategy for the hybrid cloud environment. To enable detection procedures, the proposed SDMTA mitigation architecture includes integrated network monitoring. The suggested and existing state-of-the-art models’ detection rates over the input dataset were estimated. When compared to the existing state-of-the-art model, the system's accuracy, specificity, and sensitivity were found to be 99.7%, 98.32%, and 99.92%, respectively.

Cloud computing has emerged as a new technological domain in the IT industry. Currently, several organizations working in various domains, such as healthcare, finance, manufacturing, smart grid, Internet of Things (IoT), and IT, are increasingly integrating cloud computing with their traditional applications. The key idea behind the usage of cloud computing in IoT is to increase efficiency without compromising the data quality. When it comes to collecting data of thousands or millions of servers, the cloud offers scalability and reduces the computational load on each sensor. The highly configured servers in the cloud are very useful in processing and analyzing the sensors’ data. The security of such a shared infrastructure is very crucial. It is the major barrier in the adoption of cloud-based services, followed by issues regarding compliance, privacy, and legal matters. The digitalization of critical infrastructures, such as smart grids (SGs), brings advantages and opportunities for remote access and control. It enables intelligent and online monitoring of these systems, which considerably enhances cyberattacks’ vulnerability. Cyberattacks are among the most important threats to SGs. Therefore, efficient control systems should be designed that can detect and isolate cyberattacks to keep the SG reliable and secure operation. This special section aims at providing a forum to discuss the most recent advances on security and privacy in cloud-assisted IoT and SG applications.

With the rapid evolution of the industrial Internet, cloud service has emerged as a next-generation industrial standard that has the potential to revolutionize and transform the enterprise industry. In recent years, numerous enterprises have acknowledged the benefits of cloud-based service models. However, the security issues are a major concern, such as stealthy malware attacks against virtual domains. In this article, we propose an introspection based security approach, called VMShield for securing virtual domains in a cloud based service platform, which is designed to detect malware in cloud infrastructure. VMShield performs virtual memory introspection from the hypervisor (trusted-domain) to collect the run-time behavior of processes, making it impossible for the malware to evade the security tool. The use of introspection makes the proposed approach a better choice over traditional static and dynamic state-of-the-art techniques which fail to detect stealthy attacks. The VMShield extracts the system call features using Bag of n-gram approach and selects important features using the meta-heuristic algorithm, binary particle swarm optimization. Random Forest (RF) classifier is used to classify the monitored programs into benign and malign processes, making it capable of detecting the variants of malware thus, an advantage over the typical signature-matching approach. The University of New Mexico (UNM) Dataset and Bare cloud Dataset (University of California) has been used for the demonstration and validation of VMShield. The results prove that VMShield achieves a higher attack detection rate and reduced storage compared to previously proposed techniques.