Mobile Radio Communications and 5G Networks

A neurodegenerative condition that affects the elderly is Parkinson’s disease (PD). A crucial first step in giving quick medical attention is the early diagnosis of PD. The field of artificial intelligence has recently paid increased attention to computer-assisted approaches for PD identification. The suggested method is a strong contender for identifying PD patients. The results of PD symptom monitoring using cost-effective computer tools are useful in telemedicine applications. In this paper, a model is designed to detect PD using an online dataset. Images were resized and analyzed which were classified using the Convolution Neural network (CNN). In novelty, the use of the Nearest Neighbor is used in the Pooling layer. 93% accuracy is attained using the proposed model which results in a 12.9% improvement over other state-of-the-art techniques.

So before understanding about deep learning, we should also look at Artificial Intelligence (AI) and Machine Learning (ML). The purpose of AI is to train machines in such a way that they can function like the human mind. The field of AI includes machine learning, the purpose of which is that the machine can learn by itself according to its experience and can develop such skills in which human involvement is not equal. Let us now understand what Deep Learning is. You can also say that very complex neural networks have been named deep learning, and you can also see it as an advancement in machine learning. Basic machine learning had limited data processing capabilities and generally required structured data. While the data processing capacity of deep learning algorithm is very high, and compared to traditional machine learning, it does not require structured data, rather it can handle both structured and unstructured data. In one sentence, deep learning enables computers to think, understand, and experience like humans.

With the advancement of IT innovation, mobile computing expertise has lately become more widely used by humans. Through the use of sophisticated devices like tablet PCs, smartphones, and other mobile computing devices, a comfortable atmosphere has been created. There are lots of potential threats in the world of mobile technology. Protective components are thus required to guard against safety concerns, especially the Short Text System. The harm caused by spoofing has kept rising as mobile computing environments have become more common. In this research, we examined the privacy concerns around smishing in contexts that are used in mobile computing. Additionally, we provide a strengthened security framework for identifying smishing attacks. The proposed approach enhances the Naive Bayes classification method to enhance the identification of Smishing attacks in connected phones. This framework distinguishes between legitimate text messages and fraudulent ones. The computational intelligence technique is primarily utilized to select applying information. It is therefore feasible to look into a phone message and successfully identify SMS phishing. Furthermore, we evaluate and analyze our proposed approach to show the effectiveness of the approach.

It has been used as an inspiration for much scientific and technological advancement including improvements in the field of computer science such as the Shortest Path algorithms and Environment Machine Learning. These algorithms are used in a wide range of applications, such as finding the fastest route between two cities or the shortest route for a robot to navigate through an obstacle course. One way that the Big Bang Theory has improved Shortest Path algorithms is through the development of a new algorithm called the Big Bang Shortest Path algorithm. In addition, Environment Machine Learning is a branch of machine learning that focuses on learning from the environment. This approach to machine learning has been inspired by the way that the universe and the environment have evolved over time. By using the principles of the Big Bang Theory and Environment Machine Learning, researchers have been able to develop more efficient algorithms for a variety of tasks. For example, these techniques have been used to develop more accurate weather prediction models and to improve the performance of robotic systems in complex environments Overall, the Big Bang Theory has had a significant impact on the development of computer science, and has inspired many innovative ideas and technologies.

Emotion plays a vital role in our day to day life. They can drag a human to such a state where he can feel helpless and can make some decisions which are not right for him. So, we design a machine learning model with the help of image processing and computer vision, which can detect the emotion of a person even if his face is partially visible. We are able to build a successful model with the help of CNN (Convolutional Neural Network), and in this model, we use the FER 2013 facial expression dataset with the help of which we get the validation accuracy of 72%. In this project, we combine two major projects that are facial emotion recognition and gender classification and we are able to detect the gender and emotion in live time even if the face is partially visible.

A novel strategy that is increasingly becoming popular in today's world is cloud computing. With the advent of cloud computing, the internet has evolved into a platform for computing that offers network, storage, and other resources. Additionally essential to the entire transformation of the IT business is cloud computing. It offers a variety of features, such as on-demand applications, utility-based pricing, resource sharing, etc., that draw consumers to utilize it. Regardless of these advantages, cloud is still associated with some security related risks. One main issue is the data security over cloud which is restricting the development of cloud computing. Therefore to accelerate its growth and to make it widely acceptable data security risks are to be resolved. This paper discusses that how the various security threats involved in cloud computing can be classified. It also describes the major obstacle that is restricting the development of cloud computing i.e. Data security risks and the different solutions proposed by different researchers to cater out those threats.

Paradigm and research work culture have evolved with the ontogenesis of the software development field throughout the timeframe because of their heterogeneity. Genetic Improvement (GI), which views the code as manipulable “genetic material,” is the foundation of the transplantation approach. We must capture the code in the DONOR that the selected functionality depends on in order to transplant it to an unrelated HOST. So, we espouse two eccentric frameworks in the realm of neural networks, each with its idiosyncrasy; PyTorch framework (which is notorious for deployment) and TensorFlow emphasis the domain of Artificial Intelligence. This transplantation is more than 3000 times faster than re-training. Our baseline technique took 20 times less time in contrast with ONNX Convertor. This work demonstrates that for any existing exemplar with acceptable outcomes, we don’t have to begin from scratch, nor do we need to inculcate any convertors. We also proposed the need for software transplantation and the circumstances under which this transplantation would be incredibly useful.

Investment in Gold and Silver is one among the many options for investors to earn a good amount of money. It also strengthens the national economy. However, investors need to understand the market movement and the impact of the influencing factors. The use of suitable techniques is another important aspect of adding innovation to the stock market investment process. Prior knowledge always helps in making better investment decisions and particularly it is very useful in the case of the stock market investment industry. This research aims at predicting gold and silver prices using machine learning techniques. Data on daily prices from 1 January 2022 to 20 March 2023 was analyzed using two machine learning models: Convolutional Neural Network (CNN) and Bi-Directional Long Short-Term Memory (LSTM). To maximize returns, it was necessary to determine when the best time to invest in gold and silver. From the experiments, the CNN model was found to be more effective than the Bi-Directional LSTM model.

The history of human civilization has shown that the major strength behind the advancement of technology has always been the hu- man’s need for medical and healthcare applications. For decades, cloud computing has dominated healthcare information systems based on everything as a service model. However, the long service delay offered by cloud applications is one of the significant limitations. During emergency situations, monitoring of patient’s health, decisions must be made quickly with limited available resources, as it directly affects the patients’ lives. The introduction of edge computing and fog computing as optimal technologies has solved these latency-related problems. This paper provides an open architectural framework for fog-based Internet of Healthcare (Fog-IoH) applications. Besides, various services and applications support in the integration of fog with the IoT health sector have been indicated. Further, the applicability of IoT in healthcare has been discussed by presenting a case study of smart gloves wearable devices.

Cloud Offloading is an essential method for the data decentralized distribution and management hence multiple approaches can result in intrusion. Spark MLlib intrusion detection typically uses multiple algorithms and works on Resilient Data Distribution on the cloud to detect the intrusion information, and is not equipped with the flexibility and low detection rate inability to deal with high-dimensional data, and can’t solve these issues efficiently. To enhance the effectiveness of the intrusion detection over cloud offloading, an innovative general intrusion detection framework that is being developed with the spark that has a resilient distribution of data across the cloud has been proposed in this paper. It comprises four components a preprocessing module, a label encoder module, a feedback module, and a classification module. Preprocessing module information is compressed using the module to label encoder. This creates a less-dimensional reconstruction and classification feature. The database module can store the compressed features of all traffic. This allows for retraining as it tests the classifier and then restores these features back into the original traffic. To test the framework's effectiveness, simulations were performed using CLOUD OFFLOADING DATA INTRUSION DETECTION SYSTEM CODIDS 2017 data set to match the actual network traffic. According to the test results showed, the precision of multiclass and binary classification is superior to previous research. A good level of accuracy has been achieved for assorted traffic data. In the end, the possibility of using the proposed framework for edge/fog networks is given.

COVID-19 affected each and every sector of human life. The food sector, security, employment, agriculture, etc. converged a lot during the pandemic. The coronavirus (COVID-19) pandemic has had a significant impact on the world. The virus was first identified in Wuhan, China, in December 2019, and has since spread to nearly every country. The World Health Organization declared it a pandemic on March 11, 2020. As of January 2021, the virus has infected over 89 million people and resulted in over 1.9 million deaths globally. The pandemic has had wide-ranging effects on society, including health impacts, economic consequences, and changes to daily life and social interactions. Many countries have implemented measures such as lockdowns, travel restrictions, and mask mandates in an effort to slow the spread of the virus. The impact on the healthcare system was significant due to the overwhelming number of hospitalizations and intensive care needs in some countries. The economy took a hit as well as many businesses closed and job loss became widespread. This review provides insight into the impact of the coronavirus on people and how different measures have been taken up to have a control over this pandemic.

Though there are strong laws to protect women, violence against women is increasing across the world. In this work deep learning is used for analysing video recordings to detect harmful weapons. Around the clock, women face harassment and violence. The notable uniqueness of this proposal is that Artificial Intelligence is implemented for the prediction of crime, which has never been implemented in the previous existing methodologies. Deep Learning models for image processing can detect violence with higher accuracy and thus help cops to identify the criminals. Therefore, any crime that is yet to happen is detected and the predefined contacts get an SMS so that they can know the whereabouts of the victim. The proposed method uses YOLO v3 algorithm. For higher accuracy, the dataset consists of weapons with all possible angles, merged with ImageNet dataset this objection detection algorithm was found to perform extraordinarily to detect weapons in various scenarios, shapes, and rotations. The result showed that YOLOv3 can be used as an alternative of other traditional object detection algorithms such as Faster RCNN.

Mathematical transforms gain prominence in computer vision and image processing applications. Wavelet transform provides multiresolution analysis of signals and uses adaptive basis function that was tailored to the specific signal being analysed. In this research, the lifting wavelet transform is utilized to compress medical images. Compression of medical data play vital role in minimization of storage space for Tele radiology application. The lifting wavelet transform utilizes the lifting scheme, it is computationally less complex than discrete wavelet transform. This research work proposes compression of medical images using lifting Haar wavelet transform. The achieved outcomes were found to be satisfactory, and their performance metrics were thoroughly validated.

A content-based image Retrieval (CBIR) has become an essential tool for managing and searching large-scale images. However, the accuracy and performance of CBIR systems can be improved by combining data mining techniques. Content-based retrieval (CBR) uses the properties and characteristics of the content itself to search for and retrieve information from a big database instead of depending on text or metadata. CBR is very helpful in research, where it is necessary to swiftly and effectively examine vast amounts of data. According to the results, data mining techniques can considerably increase the retrieval process accuracy and effectiveness. Similar photos can be grouped together using clustering algorithms, common patterns of visual features can be found using association rule mining, and images can be classified using classification techniques. In order to create a system for content-based image retrieval and processing, we studied the retrieval of images from huge databases using a variety of feature extraction and matching techniques. The demand for CBIR development came as a result of the sharp rise in image database volumes and their widespread use in several applications. The description of basic feature extraction methods including texture, color, and form is provided in this study. Once these features are retrieved and then used for comparing photos based on similarity. This study suggests a cutting-edge system design for CBIR system that integrates content-based picture and color analysis with data mining methods. This work is intended to develop a segmentation module for the CBIR system.

Image dehazing is a critical task in computer vision that involves removing the haze or fog from an image to enhance visibility and improve quality. Methods based on deep learning for image dehazing have demonstrated promising results in recent years, utilising the ability of neural networks for learning how to map of a blurry picture to a clear one. This study presents a thorough analysis of current deep learning methods for image dehazing. We analyze and compare these methods based on their underlying principles, strengths, and limitations. We categorize them into different approaches and evaluate their performance using standard metrics. Moreover, this paper will also explore the challenges and future directions in advancements of deep learning-based image dehazing.

Artificial intelligence has reached a stage where machines possess the capability to engage in tasks that traditionally demanded human intelligence. The integral component of this advancement lies in “machine learning,” where algorithms are trained to generate predictions or make decisions by analyzing data. In hate speech identification using machine learning, a number of methods are used to automatically find text that uses vocabulary that is considered to be derogatory, discriminatory, or motivated by hatred. Supervised learning techniques like neural networks, decision trees, and SVMs need a labelled dataset comprising samples of hate speech and non-hate speech. Unsupervised techniques, such k-means clustering, use word frequency and other variables to group comparable text data. CNNs and RNNs are examples of deep learning systems that learn complex word associations and spot trends indicating hate speech. Text data is subjected to the utilization of N-grams, word embeddings, and sentiment analysis in order to derive distinctive attributes. Accurate detection is improved by ensemble methods that combine predictions from various models. Identifying hate speech, avoiding bias in data and algorithms, and the necessity for large and diverse datasets are among the difficulties. Ultimately, machine learning-based hate speech identification is an essential tool for preventing hate speech online and fostering an inclusive and secure online environment for all users. So we did research on detecting hate speech using various algorithms. The TF-IDF representation prioritises textual terms, whereas the ensemble method uses classifier diversity to capture distinctive patterns. Results from experiments show the strategy's effectiveness, with a 90% average accuracy rate for detecting hate speech. By successfully utilizing AI's capacity to fight hate speech, this research helps the development of a diverse and secure online environment. The suggested approach works well for automatically identifying hate speech, making the internet a safer and more welcoming place for all users.

Alzheimer’s Disease (AD) and Parkinson’s Disease (PD) are all age-associated neurodegenerative diseases. The evaluation of AD and PD each is the onset of Dementia, which reasons memory impairment irreversibly, orientation, comprehension, gaining knowledge of any new things, and taking the judgment. A unifying feature of Alzheimer’s ailment (AD) and Parkinson’s ailment (PD) is the build-up and processing of mutant or damaged intra and extracellular proteins; this results in neuronal vulnerability and ailment with inside the brain. With high-overall performance computational gear and diverse deep learning (DL) strategies, those modalities have found thrilling opportunities for figuring out and diagnosing neurological disorders. The machine proposes deep learning (DL)-primarily based totally strategies to hit upon neurological disorders that specialize in Alzheimer’s ailment, Parkinson’s ailment from MRI data obtained the usage of exceptional modalities along with practical and structural MRI. The study makes use of predefined deep learning structures and fine tunes them to acquire the purpose of locating the ailment in patients with the accuracy of 99.07%, deep mastering is the subset of gadget mastering in which neural networks and numerous different forms of layers are merged collectively to shape a model to system a photograph and predict. The several DL architectures during distinctive troubles and imaging suggest the Convolution Neural Network (CNN) out performs exceptional techniques in detecting neurological disorders.

Rainfall forecasting is difficult since the weather always delays its parameters for less than a minute. Depending on one’s abilities at work, accurate predictions can aid a person and lessen asset loss. This article provides a series of experiments that employ popular machine-learning techniques to create models that forecast whether or not it will rain tomorrow based on weather information from the previous day. This comparative study emphasizes on three aspects that are Data collection, pre-processing methods, and data modelling. The findings compare multiple measures for evaluating machine learning approaches and their accuracy in predicting rainfall by examining weather data. Additionally, it uses data provided by the user to predict rainfall using the most accurate algorithm.

A successful system is required for image processing, analysis, and early detection of colon cancer, as it has become the second most serious type of cancer, affecting approximately 15% of people worldwide. Therefore, the colon cancer prediction system utilizes MRI imaging to forecast the onset of colon cancer. To effectively investigate colon cancer at its initial stages, both high and low-level characteristics are essential throughout the procedure. In this study, we propose an ensemble approach for detecting colon cancer using a multiclass colon dataset. This paper demonstrates that the suggested ensemble approach achieves an accuracy score of 98.35% and respective precision, recall, and F1-scores of 98%. Importantly, the proposed ensemble approach exhibits superior detection accuracy in comparison to existing methods, offering an improved screening approach that can significantly reduce the pathologist’s workload in identifying tumor regions. This innovative method holds the potential to serve as a robust tool for enhancing colon cancer diagnostics.

Multiplication is one of the most important and most frequently used operation in many DSP computations. Multiplier is the basic operation in many VLSI circuits. Some of the DSP applications such as convolution, correlation etc., uses the multiplication very frequently in the any design. So their performance depends on the Multiplier and the Multiplier’s performance depends on the type of adders that we used to implement it because the multiplication operations are critical so we use the adders. Multipliers can be optimized in terms of the area, delay and power by diminishing the no. of partial products produced, by using a fast adder unit and by taking care of the implementing logic. The main objective is to design a signed 8-bit multiplier which is efficient in terms of area occupied and energy by using Booth Radix-4 Encoding Technique. By exploring the characteristics of the signed numbers, the design effectively minimizes the number of arithmetic operations required, resulting in reduced latency and improved energy efficiency. Furthermore, the architecture incorporates parallelism and pipelining to enable high throughput in the multiplier module. With the extensive simulations and comparisons with the existing designs demonstrates that the proposed multiplier achieves significant energy savings without sacrificing its performance. In this proposed work the partial products are generated in parallel to improve the performance and the adders used for the partial product addition that are implemented by using decoder logic.

In order to fulfill the rising demand for electricity and decrease the use of fossil fuels, which is contributing to environmental issues, hybrid renewable energy sources (HRESs) are currently being urged to be integrated into the grid. HRES grid integration may lead to some power quality (PQ) issues. Some flexible devices should be used to reduce PQ issues and enhance the functionality of grid-connected HRESs. To minimize PQ problems such as voltage drop, swelling, disturbances, and harmonics in the hybrid power system (HPS), this research introduces a distributed power inflow controller, a form of an adaptable device. The HRES presented in this study consists of a PV system, a WT, and a BESS. This research builds an 8-bus DPFC system with real and reactive power to address the PQ problems in HRES systems. The main goal of the effort is to alleviate PQ issues and account load demand of the HRES scheme. A model predictive controlled system (MPC) and fractional-order PID (FOPID) controller are employed to drive this DPFC-PQ to determine how well better steady state analysis and MPC controller parameters can be tuned. The MPC controller and FOPID controller are contrasted as well. The outcomes demonstrated that the suggested MPC for the DFPC had reduced the PQ issues in grid attached HRESs.

The health of the mother is crucial to the well-being of the baby throughout pregnancy. Early treatments and individualized care may be more effective when maternal health hazards are properly classified. In this study, we explore the feasibility of using fundamental and boosting machine learning algorithms to categorize threats to maternal health. On an extensive collection of maternal health indicator variables, we test the efficacy of many machine learning techniques, including logistic regression, decision trees, random forest, gradient boosting, and XGBoost. With an experimental accuracy in classification of 86.48%, our data show that XGBoost performs well. To further verify XGBoost’s efficacy, we use several assessment criteria, such as the lift curve and the ROC curve. These results provide important insight into how machine learning might be used to enhance prenatal care and lessen associated hazards to mothers.

The activity of concealing information or data within a carrier entity in a way that is difficult for unauthorized users to notice is known as information hiding, sometimes known as data hiding. Different types of digital material, such as pictures, audio files, movies, text documents, and network protocols, can use information-concealing strategies. This paper proposes a text-based data-hiding approach using Huffman and the midpoint folding technique, which increases the payload capacity and reduces the embedding distortion. Results show that the average PSNR obtained is 57.23, which depicts the good quality of stego images, making them indistinguishable from the human visual system (HVS). Also, on concealing 2 KB of data in four considered dataset images, an average of PSNR = 57.23195, WPSNR = 55.31998, SSIM = 0.993238, and MSE = 0.125765, which proves the credibility of the proposed method.

The Tunnel field-effect transistors (TFETs) leverage quantum tunneling for efficient power consumption and improved switching capabilities. In this paper, we have demonstrated the impact of parametric variation on electrostatics and analog performance of the proposed tunnel field-effect transistor (ESDGTFET) device using Silvaco 2D device simulator. The paper investigates various parameters like threshold voltage, ION/IOFF ratio, drain current, subthreshold swing (SS), transconductance(gm) and cutoff frequency (fT) for different channel lengths and gate dielectric materials. The finding from the investigation reveals that the subthreshold swing (SS) is improved by 49% when the channel length is reduced from 40 to 20nm but no notable changes were observed in threshold voltage. The total capacitance of the device is also improved for a shorter channel length. Furthermore, the on current and SS of the device are improved for HfO2 gate dielectric material as compared to SiO2. For high-k dielectrics, the device’s threshold voltage drops substantially. As a result, the device functions optimally in low-power contexts.

Face-based biometric technology finds extensive use in authentication applications due to the human face being readily accessible and containing abundant information in daily life. However, facial recognition systems face the challenge of thwarting face spoofing attacks. Numerous spoof detection systems have been introduced, each with distinct merits and limitations. Ensuring precision in face-based identity recognition and access control is imperative to counter potential threats. Notable face spoofing detection algorithms, including CNNs, SVMs, modified CNNs, and various classifier techniques, have demonstrated efficacy. Nonetheless, as face spoofing tactics have evolved from exploiting printed photos to incorporating masks, there arises a pressing need for performance enhancement, necessitating regular strategy updates. This work aims to provide a comprehensive overview of anti-spoofing techniques, with a specific focus on established and widely adopted face spoofing countermeasures. The present study examines the proposed methodologies and ongoing efforts in this domain, encompassing performance assessments conducted so far. This collective analysis serves to unearth novel avenues and refine existing methodologies, contributing to the continuous advancement of this rapidly expanding realm of research.

In this research work the transfer characteristics of the n-channel double gate fin field-effect transistor (DG FinFET) are optimized for various device geometries. From the I-V characteristics of the DG FinFET structure, the off current (Ioff) is analysed. The effects of fin width, source and drain extension length, gate work function and temperature on the device characteristics are simulated to analyze the performance. The MuGFET simulation tool is employed to perform the simulations of the device. The off current (Ioff) plays, a very important role in leakage current which is one of the important short-channel effects (SCE) in nano-scale devices. In this paper, Ioff is analyzed and optimized using a different device geometry and gate material at different temperatures. It is concluded that the Ioff decreases with the higher gate material work function and lower channel width. Hence, the proposed device performs better in terms of static power dissipation in the nano regime.

Climate change's impact on hydrology is critical since it affects agriculture, vegetation, and livelihood. This investigation aims to determine whether climate change is possible in the Vaitarna River basin. The IPCC has studied temperature extremes and their effects on ecosystems (IPCC). Temperature variations are attributed to changes in solar energy, biological processes, and human activities. Climate models replicate how the climate's drivers interact using ocean, atmospheric, ice, and land surface models. Impact evaluation uses several climate models. Global Climate Models (GCMs) and GCMs are used to analyze changes in global temperature generated by doubling CO2 concentrations. In this experiment, free software SDSM 4.2 was employed. It's a blend of regression-based and stochastic weather generators in which local and large-scale data are connected. CCIS proposes the Statistical Downscaling Model as a downscaling technique. Downscaling uses GCMs to anticipate local conditions. GCMs are accurate atmospheric simulations based on large-scale spatiotemporal models. GCMs can't predict local and regional climate factors or their intensity. Dynamic downscaling didn't surpass statistical downscaling, according to a thorough study.

Among the health issues affecting women, breast cancer is a major concern. Compared to other malignancies, it has one of the higher fatality rates. Early detection of cancer patients enables medical professionals to provide a more accurate diagnosis and prognosis. The objective of the study is to evaluate the performance of the proposed CNN-based models in identifying the malignant or benign tumor types in patients based on their cancerous or non-cancerous status. The study aims to discover the important parameters that affect the model's performance during training, such as the number of convolutional layers, the quality of the training data, and the dependent variable. The study makes use of the Breast Cancer Histopathological data set that is readily available on Kaggle. It is frequently employed to evaluate CNN-based models in the healthcare sector. The paper reveals how deep learning approaches, particularly CNN models, can be used to provide robust feature representation and accurate patient predictions. The parameters for estimation performance were, in order, 97.39% precision, 97.42% accuracy, and 97.45% recall. The results of the study lend credence to the notion that applying deep learning techniques can assist doctors in making precise diagnoses, picking the most effective course of treatment, and keeping track of patients’ prognoses. It provides clinicians with a solution that is far better than standard practices. According to the study, using machine learning and deep learning techniques may greatly improve the management and interpretation of healthcare data.

In this paper, recommender systems (RS) are examined to make numerous improvements in recommendations based on Sentiment Analysis (Asani et al. Mach Learn Appl 6:100–114, 2021). The proposed approach includes an examination of each group's interpersonal relationships and personality make-up to increase the precision of the grouping recommendations (Haruna, Appl Sci 7(12):1–25, Singh et al., Int J Bus Syst Res 15:14, 2021). In this way, researchers can more accurately imitate the discussion process in which group of people engage in when deciding on a shared activity. It is also taken into consideration how they anticipate the system in a long-term recommendation process. Major consideration is on finding influencing factors and techniques for recommender systems using sentimental analysis using existing information on the recommender system and sentimental analysis (Singh et al., Int J Bus Syst Res 15:14–52, 2021). It is accomplished by including a collection of previous recommendations, which raises user satisfaction among anyone whose preferences weren't taken into account in earlier recommendations of social networking (SN). The determination of influencing factors and techniques for improving recommendations on the basis of sentimental analysis is the main goal of this paper.

Enterprise Benefits are provided by blockchains like Hyperledger Fabric, Corda, and Quorum. For instance, Hyperledger Fabric adapts to your requirements whether you work with healthcare products, real estate, or the financial industry. To digitize the recording of contracts between known parties, Corda was developed. Quorum, on the other hand, was created for the financial services industry in order to ensure transaction and contract confidentiality. In order to determine which of these corporate Blockchains technology is appropriate to implement into their enterprises, many blockchain users (businesses or people) distinguish between them. In this review paper, we provide a thorough review of enterprise blockchain technology, comparing Hyperledger Fabric/Corda and Quorom to determine which will work best for enterprise businesses.

Breast Cancer (BC) is a type of intrusive illness and considered as the most prevalent disease for women that is responsible for huge number of deaths in the world. An early prognosis could boost the treatment’s efficiency and depress the fatality rate largely. Computer-Aided Diagnosis-(CAD) relies heavily on the automatic classification of BC utilizing histopathology images; however, the accuracy of the feature-based classification method is usually highly based on the accurate cell segmentation and feature extractions. Because of impurities, overlapping cells, and uneven radiation; feature extraction and accurate segmentation are still difficult. Contrarily, the existing system possesses some significant drawbacks namely; poor choice of features, and classification efficiency, computational complicatedness, imbalanced datasets, and absence of data augmentation. To overcome these difficulties, an efficient BC classification model which includes Double-Shot Transfer Learning (DSTL) and Cross-over Tanh-based Extreme Learning Machine was proposed. The proposed model was simulated tested using the BreaKHis dataset. The accuracy of about 96.67% was obtained which demonstrates the efficiency of the model.

Non-Orthogonal Multiple Access (NOMA) is a communication methodology used in modern cellular networks that has higher spectral efficiency than the traditional schemes like TDMA, OFDMA, CDMA etc. by enabling more than one user can communicate simultaneously. Sometimes, direct transmission among the transmitter and receiver is not possible because the distance between them is greater than the transmitting range of the source. This decreases the network capacity and coverage. So, to overcome this problem this article proposes a relay supported network which can improve the network capacity by reducing the distance between the source and destination. Simulation results demonstrate that the proposed scheme has the highest achievable sum rate of the existing conventional scheme.

A brain tumour is caused by a mass of random cells inside the brain, which is dangerous and harmful to the brain. Today, it is difficult to accurately recognise brain images. The current demand is for research on the storage, processing, and manipulation of medical image data utilising contemporary technology with little human involvement. The study and creation of cutting-edge technology for the analysis, representation, and interpretation of visual data are known as image processing. The requirement for an effective storage model that may assist in saving the brain MRI images is investigated in this study with the use of an image processing approach. A matrix-based technique is suggested to store the brain MRI pictures with less storage capacity. This model converts DICOM-formatted brain MRI images into matrix format. The patient’s information and image data are combined in the DICOM images’ image data and header information. These information are transformed and put in the matrix. The suggested model’s input is obtained from the stored matrix. The suggested model employs many stages of image processing. The procedure begins with pre-processing of the brain MRI pictures, then clusters the white matter (WM), grey matter (GM), and cerebrospinal fluid (CSF), segments the tumour, and categorises the tumour before handling storage of the MRI images. Filtering methods are used on MRI data at the pre-processing stage to get rid of noise and text artefacts. The K-means clustering technique is used to separate the white matter, grey matter, and CSF. With regard to both pixels and edges, the MRI picture is segmented. When it comes to tumour classification, RBF kernel-based SVM ensemble performed better, and SVM ensemble is used to classify brain tumours.

The design of U-slot microstrip patch antennas is the objective of this paper. This U-slot patch antenna can produce dual band operation. A configuration model was utilized for the purposes of designing, simulating, and analysing a microstrip patch antenna with a U-slot loaded over two dielectric layers. Miniaturization is accomplished with the help of RIS, which stands for reactive impedance surfaces. Simulation parameters are employed in the process of determining antenna framework such as Gain, Polarization, Return Loss, and Impedance. The performance of the suggested U-slot, RIS-based antenna was superior to that of the regular patch antenna and the patch antenna combined with the U-slot. Applications that make use of Bluetooth are a good fit for the proposed antenna. This antenna is mainly used for ISM band for required bandwidth and proper applications. Microstrip patch antenna with reactive impedance surfaces as in results is providing better results than normal patch antenna.

Heart disease is amongst the key contributors to increasing mortality globally. Heart diseases affect many people of middle or old age, causing adverse effects such as stroke and heart attack in many cases. Therefore, effective diagnosis and diagnosis of heart disease is essential for prevention of serious health problems in the present scenario. In HDP (heart disease prediction), the futuristic probability of the coronary disease is forecasted based on the existing information. The heart disorder is predicted in diverse stages such as to pre-process the data, extract the attributes, and classify the data. This work concentrates on reviewing numerous techniques to predict the coronary disorders. The algorithms used for this work include Logistic Regression, Naïve Bayes, Support Vector Machine, K-Nearest Neighbour, Decision Tree, Random Forest, XGBoost. In results Random Forest and XGBoost turn out to be the most effective algorithms, achieving an accuracy of 86.89% and 78.69%, respectively.

Since mobile devices have limited resources, they can offload computationally heavy tasks to the edge of the network via a mobile edge device. Recent years have seen a lot of focus on the development of an optimal offloading approach for deciding whether a job should be completed locally or at an edge device. Systems as they stand are inadequate for practical usage because they were primarily developed for scenarios involving a single mobile device or a single edge device. This work presents offloading rules that minimize task drop rates and execution delays in scenarios where there are many non-cooperative mobile devices and multiple edge devices, without requiring knowledge of the environment's dynamics. Mobile devices are only able to partially satisfy this non-cooperative resource allocation dilemma. We show how a reinforcement learning-based strategy can be used to reduce the effects of partial observability and gradually understand the dynamics of the environment and its long-term repercussions. The suggested algorithm dramatically decreases task drop rates while minimizing energy and computing costs, making it preferable to preexisting offloading schemes.

Mobile networks of the fifth generation (5G), which will consume less energy and offer better quality of service (QoS), are anticipated to accommodate the enormous volume of data traffic. The key enabling technologies like mmWave methods, large multiple-input multiple-output networks, and heterogeneous networks (MIMO) are highlighted to achieve this. Regardless of the technology employed, a user association mechanism is required to determine if a user is linked to a given base station (BS) prior to transfer. User Association is essential for improving the network's load balancing, spectrum utilization, and energy efficiency. The subsequent significant development in mobile communications standards is 5G, which will follow the next 4G standards. Due to 5G technology, most high-bandwidth consumers’ access to their phones will change. To better utilize the services, User Association enables the mapping of User Equipment to base stations. However, the channel fluctuations in a mmWave system might be quick and unexpected, potentially making centralized user association ineffective. An effective distributed matching algorithm must be created and customized for user association in 5G Heterogeneous networks.

Smart glasses for blind are a promising technology that can help blind people to read and interact with the real world in a very convenient manner. This paper will explain the development of the text-to-speech prototype. This project of ours is able to convert any text into audible speech which is transmitted to the earphones allowing the blind person hear the text conversion. In this paper, we have discussed various components of our prototype. For example, text identifying libraries, camera modules, speech synthesizing models and basic hardware are required to run all this in an efficient way. We have also carried out a survey with a small number of blind people for analysis of the prototype. According to the feedback, we also made changes in the prototype.

False news detection has become a more important research area as fake news can be harmful to your mental health and makes it harder for people to see the truth. The purpose of this research is to shed light on the issue from the standpoint of foundational linguistic management to provide the groundwork for the detection of misdirection in the media. The fundamental challenge in this field of study is the lack of high-quality data, since it may include both bogus and real reports on a representative sample of the population. In addition to the previously used malleable and dominant truth disclosure structure, a truth acknowledgment approach based on the concept of nearby words has been discussed. The use of almost identical terminology and phrases may help expose controlled fake news. The Jaccard estimate is utilized in the primary computation to determine the rational number of fake news stories with a firm quality score, using the same key phrases. Scores for freedom, mien, and weakness are averaged to arrive at a measure of stability. The SRTD programs are more accurate than other truth identification techniques. The wordnet dictionary can enhance the accuracy when applied with SRTD.

Grapevine diseases may possess a substantial effect on crop yield and quality, but early identification and mitigation can help prevent losses. The manual detection and diagnosis of vineyard diseases can be time-consuming, subjective, and difficult for producers who lack the requisite knowledge. Therefore, automated systems may offer a more effective and accurate method for classifying grapevine diseases. This paper recommends an investigation on Grapevine Leaf Diseases Classification with Deep Learning Techniques and Feature Extraction Using the Internet of Things. The purpose of this research is to develop an IoT-based system for identifying and categorising different kinds of grapevine leaf diseases in real time. The system will capture photos of grapevine leaves using a network of IoT devices outfitted with high-resolution cameras. The images are going to be transmitted to a cloud-based infrastructure for analysis using techniques for deep learning and the extraction of features. The system that is suggested will enable cultivators to detect and categorise grapevine leaf illnesses in real time, enabling them to take swift action to prevent the disease’s spread and increase crop yield. The research will utilise the Grapevine foliage image dataset accessible via Kaggle and other datasets that are freely accessible to train and evaluate deep learning models. The models will be fine-tuned and optimised for the IoT platform in order to process and analyse images in real-time. The efficacy of the system will be evaluated using the metrics of precision, recall, precision, accuracy, and F1-score. The proposed research has the potential to revolutionise the grapevine field by providing producers with an affordable and effective instrument for detecting and classifying grapevine leaf diseases. In addition to being extensible to other crop diseases, the IoT-based system can contribute to developing sustainable agricultural practices.

The Internet of Things (IoT) is a term that refers to the ability of objects to interact with one another via the use of intelligent connecting devices. Conventional Vehicular Adhoc Networks (VANET) were able to upgrade. Due to this functionality, they have renamed the Internet of Vehicles (IoV). Real-time data exchange between automobiles, sensors, vehicles, roads, and personal gadgets via wireless communication is referred to as the IoV. The differences between VANETs and IoV are frequently misunderstood, leading to ambiguities. This paper elaborates on the differences between the communications architectures of VANET and IoV. In addition to that, the protocols related to the VANET and IoV are discussed and evaluated. The performance of the three different protocols, Greedy Perimeter Stateless Routing (GPSR), Traffic Dynamism-Balanced Routing Protocol (TDBRP), and Intersection Gateway and Connectivity based Routing (IGCR), are evaluated on the different parameters like Packet Delivery Ratio (PDR), End-to-End (E2E) delay and packet drop ratio. Simulation results exhibit that IGCR shows a better PDR and TDBRP shows the minimum E2E delay.

Optical Character Recognition (OCR) systems are computer programmers that read text from scanned documents and images. Character recognition and text detection are two texts are analyzed. Characters are classified according to their pattern descriptions or features in the categorization process. The characters are identified using a particular classifier. The unified character descriptor (UCD) to be proposed for characters of the attributes in this environment. After, the matching is used to check that the classification was correct. The appreciation strategy functions well know similar scanned documents, but it cannot distinguish characters with a lot of font distortion and variation. Classifiers based on deep neural networks (DNN) could be used to enhance recognition. The MLP (multilayer perceptron) ensures a high level of recognition when providing thorough training; precision is essential. In addition, the convolutional neural network (CNN) is increasing in popularity because of its strong performance; it has gained. Furthermore, MLP and CNN may both be affected by the training process. We make a comparison amongst MLPs in this study as well as CNN. MLP receives the UCD description as well as the necessary network configuration. We worked for CNN to use a convolutional network developed to recognize machine-printed characters and handwritten (Lenet-5). We modify it to accommodate 62 different classes, including characters and digits. Furthermore, graphic processing unit (GPU) parallelization is examined to speed up both CNN and MLP classifiers on our experiments; we show that while classifying characters, the employed real-time MLP is 2xless relevant than CNN.

With advancements in sensor technology, classification of hyperspectral image (HSI) has gained popularity in the field of research. The graph convolutional neural network (GCNN) approach is widely used for applications such as learning over graph representation and semi-supervised learning. Research studies demonstrate learning results of graph filers and CRNN having superior results compared to other linear models in extracting hyperspectral image features. However, when applied on real world pattern learning situations CRNN performance have shortcomings. In this study, deep hybrid-multi-graph neural network (DHMN) approach is used to facilitate pattern learning using convolutional network approach. To solve issues related to (i) smoothing-spectral filter approach is used to extract spectral features and (ii) autoregressive moving average (ARMA) filter approach is used to prevent noising. HSI datasets are used to conduct experiments. To smooth and refine deep hybrid network features, GraphSAGE-based network approach is used. Results of experiments demonstrate DHMN performance is better than CRNN model.

As the amount of human–computer connection has increased steadily over the past few years and emotion recognition in voice has attracted a lot of attention. This study uses DNN (deep neural networks) and MFCC (mel-frequency cepstral coefficients) to provide a novel method for audio emotion identification. The suggested approach seeks to precisely categorise speakers’ emotional states based on their auditory signals. The features from the voice samples are retrieved using MFCC, and the deep neural network model uses these characteristics as input. To associate the retrieved MFCC characteristics with the appropriate emotional labels, this model was trained using a sizable dataset of speech samples that had been categorised and covered a wide spectrum of emotions. This concept is used to capture discriminative features for voice sentiments recognition and deep neural networks to achieve state-of-the-art performance, outperforming traditional machine learning methods in terms of accuracy, precision, recall, and F1-score with RAVDESS (Ryerson Audio-Visual Database of Emotional Speech and Song). The outcomes show how well the MFCC features work in extracting emotion-related data from audio signals. Additionally, the DNN obtains an average performance of 61%, demonstrating its capacity to effectively learn and categorise emotions.

The research’s main goal is to identify the crucial elements of putting machine learning techniques into practice in order to improve electronic commerce (E-business) inside the company. The data gathered will be analyzed to achieve this. In order to improve e-business and achieve sustainable development, researchers have concentrated their efforts on taking into account significant machine learning determinants that affect demand forecasting, the application of ML in purchase behavior, improving customer engagement, and support in overall cross-selling of the products. The major objective of this essay is to provide an in-depth and critical understanding of how machine learning technology may be used to enhance the performance of online businesses. As a result of the dynamic nature of the current environment in which businesses operate, enhance the value of their brands. Because of this, having an understanding of the machine learning approaches that the owners of these websites have used is more significant in this research. This is because management needs to be able to support their E-business and create the most money feasible for the owners of these websites. Because they use this kind of website approach, these commercial websites are better able to withstand the intense competition that exists in their field. In this brief message, we present these cutting-edge websites that make use of deep learning-driven machine learning and artificial intelligence. These websites are dynamic and cutting-edge at the same time.

A WSN is a wireless network comprising small sensor nodes to monitor environmental or physical parameters. Since wireless sensor networks have limited computational power, memory, throughput, and energy, traditional security solutions designed for resource-rich systems are inappropriate. Given these limitations, providing basic security methods for data transfer in wireless sensor networks is vital. Our work is divided in three phases. Phase 1 focuses on development of a pairwise key management technique. We suggested broadcast tree construction for a wireless sensor network in phase 2. In phase 3, we proposed an enhanced watchdog strategy as a practical way of detecting rogue nodes. The main objective of this model’s depiction is to emphasize how important it is to reduce network power consumption to prolong network lifespan and identify and terminate rogue nodes before they broadcast packets. Experimental analysis shows that our model provides better results than state of art systems.

Educational failure is prevalent. The surge in the number of students who quit school has multiple root factors. The inability to succeed academically is a primary factor in why students drop out of school. Since many students struggle to adjust to their new school, this affects performance. Our study aims to identify all elements affecting undergraduate academic achievement. Thus, this initiative aims to help students identify the factors that lead to their successes so they can take steps to change their results. Students, course instructors, and others can improve the environment after identifying and assessing its main components. We used a Recurrent Neural Network and Long Short-Term Memory classification technique to forecast student academic success early. This method is compared to numerous machine learning classifiers and a deep learning classification model. Using study findings from numerous trials, we examined the classification performance of several standard machine learning techniques, such as support vector machine, random forest, J48, artificial neural network, and naive bayes, as well as deep learning models, such as RNN. RNN-LSTM sigmoid, Tan–h, and ReLU function are used to predict student performance and enhance teaching. The results are compared to deep learning and machine learning methods. RNN-LSTM (ReLU) has the highest accuracy rate of 97%, as per experiments. Our technique has great classification accuracy on different datasets or real-time complex huge datasets of students with multivalued variables.

The use of data mining for the task of analyzing behavior in the context of online educational settings is especially well suited for its application. This is due to the fact that data mining can analyze data and unearths hidden truths that are disguised within the data itself, while doing so manually would be difficult and time-consuming to do. This is as a result of the fact that data mining carries with it the prospect of unearthing information that is concealed within the data itself. A significant number of companies are now enhancing both their understanding of the industry as well as their ability for decision-making by using various data mining tools and methodologies. Educational institutions are increasingly turning to data mining techniques in order to improve their infrastructure, increase their student retention rates, and improve their average grade point averages. This article explores the numerous applications of educational data mining, with a particular focus on the ways in which it may be used in online and other forms of distance education. The use of data mining (DM) in a variety of educational settings is the primary purpose of educational data mining (EDM), a multidisciplinary field of study that is developing at a rapid rate. The development of methodologies for the analysis of certain kinds of data derived from educational settings is the primary purpose of this project. Educational information systems are able to hold vast volumes of data because of their capacity. This data may originate from a broad variety of sources, be stored in a wide variety of file formats, and be broken down into a wide variety of granularities. Because every educational challenge has its own distinct objective and collection of characteristics, it is essential to adopt a unique strategy for tackling each one if one is to be successful in overcoming the challenges they provide. Because there are so many different types of data and so many complications involved, it is difficult to use classic DM procedures in a straightforward way.

Clustering is an efficient approach for boosting network durability, energy efficiency, and sensor node connectivity in a wireless sensor network (WSN). The WSN routing protocol has been thoroughly investigated. Based on the network organization, the wireless sensor network has been classified into three types: flat routing, location-based routing, and hierarchical cluster-based routing. Cluster-based routing, because of certain advantages is more efficient in routing technology. This research paper presents the outcome of a large scale survey that was done on cluster-based routing algorithms in WSN. Finally, the authors discuss the highlights and challenges of clustered routing techniques.

Stroke has a significantly detrimental impact on the human body which necessitates rapid medical care and treatment. There have been widespread initiatives to enhance stroke detection and treatment in response to the overall costs associated with stroke. Early identification and appropriate therapy are essential for minimizing risk to the affected region of the brain and preventing secondary problems. A supportive decision-making model has been designed in this study to analyze the electronic health records of the patients. The automated model is supplied with an extensive set of health and lifestyle factors from the patient’s database in order to make the necessary stroke likelihood prediction. This predictive analysis can reveal the major and interconnected health factors which can raise the probability of stroke. A two-phase designing process is employed to construct the automated prediction. In the first phase, different well-known tree-based models namely Decision Tree, Random Forest, Extra Trees, Gradient Boosting, and AdaBoost are implemented with necessary hyper-parameter adjustments. Random forest is the superior model which has shown the best possible outcome in this phase. The comparative analysis drawn among the employed models highlights the enhanced predictive accuracy of 0.9482 as exhibited by the Random Forest model. However, the f1-score of this model is substantially inferior. Hence, in the second phase, a k-fold cross-validation is further applied to this model to improvise the predictive efficiency of the random forest model. Finally, a promising efficiency is delivered by the cross-validation method with an accuracy of 0.95323 and F1-score of 0.96161.

Machine Learning plays a critical role in automating tasks that require human wisdom. Although significant studies have been done on segmenting brain tumors using multiple categories of Magnetic Resonance Images, a human evaluator is still required to bridge the semantic gap between the findings of machine learning approaches and those of the human evaluator. The inconsistencies in the shape of the Brain Tumor, as well as its many types and subtypes with varying behaviors, are the reasons behind this. Various strategies for segmenting a brain tumor and extracting characteristics are deeply analyzed in this paper.

There have been numerous developments in distributed and parallel computing over the past decade. A huge quantity of data is created every day from a number of sources, and the increasing dissemination of data has led to the creation of several frameworks capable of efficiently managing such massive data. Apache Hadoop is a software component of Google MapReduce that has attracted the interest of numerous researchers. For improved performance, proper scheduling of tasks is necessary. Many efforts have gone into the creation of existing MapReduce schedulers as well as the development of new efficient procedures. The assessment of various scheduling algorithms has been presented. In addition, we classify these algorithms based on a variety of quality parameters that influence MapReduce performance.

Healthcare is a significant domain for continuous inventions and sustainable developments over decades in majority of the countries. Due to rapid technological advancement in engineering now healthcare can’t be imagined without technologies like Artificial Intelligence, Cloud Computing, Data Analytics, Machine Learning, etc., for better prediction, analysis and to take precautions on early stages of any disease or health issues. The advent Internet of Things (IoT) spreading its feet from last one decade in healthcare industry which all together consumes the benefits of the above-mentioned technologies for the betterment of healthcare services to consumers. In this paper, the focus is especially on Internet of Medical Things (IoMT) its applications, services, and challenges in healthcare.

Cardiac disease remains the primary cause of mortality on a global scale, highlighting the urgent requirement for precise and dependable classification methods in the identification and management of cardiac conditions. The imbalanced distribution of data within cardiac disease datasets, on the other hand, poses a significant challenge to the accuracy and effectiveness of classification models. Oversampling techniques have emerged in recent years as a promising approach to addressing this issue and improving classification performance. This study focuses on analyzing the impact of one such oversampling method, SMOTE, on the classification of cardiac diseases. We investigate the effects of combining SMOTE with commonly used classifiers such as SVM, KNN, DT, RF, and NB. Further, we use the Cleveland (DCL), Hungarian (DHN), and combined datasets (DCM) from the UCI repository, to conduct our experiments. The experimental results show that using SMOTE improves classification accuracy significantly. The average performance accuracy improvement of SVM is 6.65%, for KNN is 2.88%, for DT is 7.86%, for NB is 6.78%, and for RF is 15.08% across the datasets studied. These findings highlight SMOTE's efficacy in addressing the class imbalance challenge in cardiac disease datasets, resulting in improved classification accuracy.

Without cloud computing, technology today would not function. Task planning and resource distribution play a significant role in cloud computing. This research suggests utilizing a hybrid strategy to schedule tasks and distribute resources. Each job in this manner goes through an MAHP process before being allocated to cloud resources. The suggested system also employs LEPT preemption to stop resource-intensive tasks. When measured by turnaround time and response time, respectively, experimental results show that the suggested system performs better than the current BATS and improved differential evolution algorithm (IDEA) frameworks, demonstrating that the divide-and-conquer strategy is responsible for the improvement.

Manual feature engineering can take a long time and be ineffective at capturing complicated patterns, while choosing the wrong features can produce less-than-ideal outcomes. As a result, effective feature engineering and selection strategies are crucial for enhancing machine learning evaluation. To improve the assessment of machine learning algorithms, we suggest intelligent feature engineering and feature selection strategies in this study. The two key phases of our strategy are feature engineering and feature selection. We use cutting-edge techniques like deep learning and autoencoders for feature engineering to automatically extract pertinent representations from raw data. High-level characteristics that represent intricate linkages and buried patterns can be extracted using these techniques. We use sophisticated algorithms, such as statistical methods, evolutionary algorithms, and correlation analysis, during the feature selection step to find the most informative features while minimizing dimensionality. Results from experiments show that models created with our intelligent feature engineering and selection strategies perform better than models created using more conventional methods, with an MSE value of 0.0202920. We intend to investigate novel deep learning architectures created expressly for feature engineering tasks in upcoming research. We also intend to research dynamic feature set adaptation techniques for feature selection based on reinforcement learning. For further improvement, ensemble methods integrating various feature engineering and selection strategies will be investigated. Additionally, we stress the necessity of standardized benchmarks and evaluation procedures to enable fair comparisons between various feature engineering and selection methods and to promote improvements in the field.

In order to understand consumer behavior and inform marketing tactics for Bellabeat, a high-tech manufacturer of health-oriented products for women, the article analyzes data on smart device usage. The study makes use of Fitbit Fitness Tracker data that was gathered over the course of a month in 2016, with a particular emphasis on variables relating to exercise, sleep, and heart rate. In the study's background, it is said that Bellabeat hopes to increase its market share in the global smart device industry by analyzing data from smart devices. The study emphasizes how crucial it is to comprehend consumer behavior and usage patterns in order to properly direct marketing initiatives. Data preparation and cleaning using RStudio, exploratory data analysis, statistical summaries, and data visualization approaches are among the strategies used in the analysis. According to the report, users’ daily activity levels are trending upward, with variable levels of intensity throughout the day. It shows the average distance users travel at various intensities of activity, showing that users go farther at low intensities than at higher ones. The data also shows a link between daily steps taken and calories expended, lending credence to the idea that higher levels of physical activity result in higher calorie expenditure.

Consumption and generation of renewable energy play pivotal roles in global energy dynamics, with an increasing emphasis on eco-friendly and sustainable practices. This research paper aims to examine the trends in energy consumption and renewable energy generation in selected countries beginning with an analysis of the historical trends in energy consumption for major economies such as China, the United States, India, and others; the research investigates the future of energy consumption. The paper investigates the growth rates of energy consumption over time and identifies specific years that deviated significantly from the overall trend. In addition, it examines the impact of various economic alliances, such as BRICS and OECD, on the landscape of energy consumption. In order to improve the precision of predictions and forecasts, this study employs a rigorous methodology that involves manual hyperparameter optimization. In conjunction with this, we utilize the predictive potential of several advanced regression techniques based on machine learning. To model and forecast the time series energy consumption data, specifically, Lasso regression and tree-based gradient boosting regressors are used. Through meticulous calibration of hyperparameters, we optimize the performance of our predictive models, thereby ensuring superior accuracy and robustness in predicting future energy values. Lasso regression facilitates feature selection and regularization, thereby minimizing overfitting and augmenting model generalizability. In the meantime, tree-based gradient boosting regressors utilize ensemble algorithms to capture complex nonlinear relationships within time series data. The findings of this study provide valuable insights into the historical energy consumption patterns of the world’s leading economies and the ascendance of renewable energy sources.

The neural network is proposed for the classification of six different gases (ammonia, acetaldehyde, acetone, ethylene, ethanol and toluene). The proposed network is trained, validated and tested on the 13,910 dataset points having 129 feature sets of parameters. These parameters are generated from commercial chemical sensors at different concentration levels. Initially, the database is analysed to compute the correlation between data points. The proposed neural network has three layers with different activation functions and different number of nodes. The input layer has 32 nodes, the hidden layer has 16 nodes and the output layer has 6 nodes. The network is trained in a batch size of 8 data points for 70 epochs. The model loss and accuracy for the training and testing phases are plotted. Finally, the confusion matrix of six gases is presented in the paper. The 746, 808, 447, 568, 872 and 524 data points of gas ammonia, acetaldehyde, acetone, ethylene, ethanol and toluene respectively are classified correctly. The proposed model can be enhanced and trained for other similar gases using a transfer learning approach.

This paper presents a theoretical scheme of Bidirectional Controlled Quantum Teleportation of a one-qubit state by utilizing nine qubits as the quantum channel. In the proposed scheme, Alice and Bob symmetrically teleport quantum information to each other under controller Candy. Alice transfers an unknown 2-one qubit state to Bob and simultaneously Bob teleports an unknown different 2-one qubit state to Alice under the permission of Candy as a controller. The quantum state information can be successfully transmitted if Alice and Bob apply Bell-State measurement and controller Candy applies Single-Qubit measurement on their qubits. By using the classical channel, users applied an appropriate unitary transformation on their qubit to reconstruct the desired state. The proposed protocol is run on ibmq_qasm_simulator of 32 qubits on the IBM Quantum Composer platform. Then, we take a comparison of our scheme with other schemes on the basis of the type of protocol, quantum information transmitted, classical resource consumption, quantum resource consumption and intrinsic efficiency.