International Conference on Innovative Computing and Communications

Nowadays various applications are available for plant disease identification using images. Early-stage disease identification can reduce losses and cost in cultivation. Efficient image segmentation is required to improve the performance of plant disease identification. Selecting appropriate segmentation techniques to extract an accurate object of interest while preserving original image properties is a challenging task. This paper presents the principle of image segmentation covering different techniques from traditional thresholding to the latest convolutional neural network-based approach. We reviewed the selected paper based on image segmentation and crop disease identification. Various algorithms are grouped based on the working principles like edge-based, region-based, and combining both properties. Performance evaluation of these algorithms was carried out using factors like time required, accuracy, and similarity to the original image. Holistic image segmentation based on convolutional neural network, K means clustering, etc. algorithms applied to real field crop images. The grab cut algorithm proves very useful for real field crop image segmentation as it preserves original image properties. Combining region and boundary-based techniques and automating segmentation need to be explored in future research work.

Social media is invariably being used these days for exchanging information and views on global affairs including COVID-19 pandemic. In this study, we have worked to understand the public attitudes of people in different countries towards COVID-19 vaccines using social media platform Twitter. We have applied natural language processing techniques of sentiment analysis to get an insightful outlook on people’s views. Hence, we categorized our results into fine-grained polarities to grasp the exact sentiment. For analyzing the sentiments, we have taken tweets that expressed sentiments for all countries, as well as for four countries that had higher fatality rates are United States of America, Mexico, Brazil and India. The people have expressed a neutral opinion towards the vaccines. Based on the sentiment, the vaccines were also ranked in which the people have expressed more faith in Sputnik V and Covishield vaccines.

Through the years, plant diseases have been a consistent risk to food security. Hence, their rapid identification could significantly mitigate the economic losses around the world, also reducing the harmful effect of manures and pesticides on the climate. When the disease is recognized, matching the characteristic trait, appropriate supervision measures can be applied. The idea of Precision Agriculture provides well-timed automation of agricultural processes by applying the methods of computational engineering in the agronomical domain, machine learning being the most researched and deployed technology. Furthermore, the authors have implemented a customized siamese neural network (SNN) for the originally collected tomato leaves dataset of 155 images with the achieved accuracy of 83.749% and 80.4% for training and the testing set respectively.

Images have been a rich source of information in recent years. Images are available in vast quantities, and most solutions necessitate real-time picture processing. This necessitates the creation of images with human-like capabilities for detecting and locating items in images. Object Detection is a branch of Computer Vision that has applications in a variety of domains, including Face Detection, Video Surveillance, Autonomous Driving Cars, and Medical Image Processing. Object detection should be quick and accurate. For accurate detection, all portions of the image should be searched for objects of all types and sizes. This necessitates a large computation cost as well as a significant quantity of time. It is natural and requires little effort for people, and researchers aim to create models that behave similarly to humans.

The Internet of things encompasses many devices connected to the Internet including wearable devices or edge devices forming a smart system to manage many tasks without the intervention of humans. Securing such systems, whether it is a part of IoT framework or any industrial software, is a very crucial task to be considered. Protecting them with passwords or even encrypting them still leaves some loopholes in the system. The attacker uses the concept of brute force to predict passwords and hack the system. To protect and prevent the system, many policies related to passwords are adopted. This includes continuous updating of passwords, which creates another stress to memorize them. This paper depicts a novel concept of a bombing algorithm to create a secure system and help remove unnecessary stress. The further section describes the related work done and the methodology adopted. The last section displays the results and evaluation of the proposed system.

Cloud computing is an era idea wherein customers use far-off servers to keep statistics and applications. Cloud computing sources are demand-pushed and are used inside the shape of digital machines (VMs) to facilitate complicated tasks. Deploying a digital system is the method of mapping a digital system to a bodily system. This is an energetic study subject matter, and numerous techniques have been followed to cope with this difficulty inside the literature. Virtual system migration takes a sure quantity of time, consumes plenty of sources, influences the conduct of different digital machines at the server, and degrades machine performance. If you've got got a massive variety of digital system migrations on your cloud computing machine, you may now no longer be capable of meet your provider stage contracts. Therefore, the maximum trustworthy manner to lessen statistics middle electricity intake is to optimize the preliminary placement of digital machines. In the deployment method, many researchers use ant colony optimization (ACO) to save you immoderate electricity intake discounts. This is because of its effective comments mechanism and allotted retrieval method. This article information the contemporary techniques for digital system positioning and integration that let you use ACOs to enhance the electrical performance of your cloud statistics centers. The assessment among the techniques supplied here exhibits the value, limitations, and guidelines for improving different techniques alongside the manner.

Many times, when people travel on the road, especially through rain-affected cities, it becomes a constant need to monitor road quality to essentially make sure that there are no unforeseen circumstances. Owing to the gravitation of human needs toward a smart city and the decrepit road infrastructure, the aim of this research is to constantly monitor the road surface to improve the quality and ensure that the car journey is safe. A mobile application has been developed which collects data from the in-built sensors of the phone. The collected data is sent to a database built using Google’s cloud platform, using a socket connection. The socket connection enables the data to be sent in real time. Google’s Firebase Database offers a real-time database, which can be deemed appropriate for this purpose. The analysis involves fetching the stored data, the fetched data is further cleaned and converted into a suitable form. The Z-Thresh algorithm was used to accurately determine the location of the potholes as it considers the minimum value of the z-axis accelerometer as a threshold. The analysis is done over a cloud-based tool, which further informs the authorities about the locations of potholes via SMS so that they can fix them.

In this work, we propose to learn the fusion process between the dominant skeletal features and the RGB features. This is in contrast to the previous fusion methods that simply fused these multimodal features, without learning the fusion process to exploit the semantic relationship between them. Here, we propose a gated feature fusion (GFF) of multimodal feature data which provides attention to the appearance stream of RGB data using the temporal skeletal data. Initially, the features from RGB and skeletal frames are extracted using CNN models. Subsequently, the gated fusion network fuses the features from pose and appearance domains using temporal convolutions which are further combined into a latent subspace. Finally, the latent subspace features are classified using fully connected layers with the combined loss embeddings. The proposed architecture has performed better than the state-of-the-art models on RGB-D action datasets.

With the advent of urbanization, the introduction of smart cities is taking place at a rapid rate to enable the ever-growing population in the urban cities to give them a chance of having a good lifestyle. Smart cities aim to do so by using and adopting the modern concepts of technology. The objective of this research is to understand and unleash how the smart cities that are coming up depend on technological aspects like sensors and actuators so that large volumes of data can be both stored as well as utilized to extract information that could prove to be beneficial for the growth of the city. The study followed content and document reviews with a systematic literature review to arrive at the observation made. For conducting this study, secondary data has been taken into consideration where the database of reliable sources like EBSCO, Scopus, and Web of Science have been utilized. This study has shown that how the emergence of ML (Machine Learning) tools makes use of algorithms that help in providing personalized services as well as efficient resource management in smart cities.

With the increase in research and development in communication technology, it is predicted that more and more number of sensing devices will be added in various sectors by application of IoT. Therefore, there is an immediate need of replacing the traditional methods of storing, sorting and sharing of data that has been collected from various sensing devices (Chiang and Zhang in IEEE Internet Things J 3:854–864, 2016), (Lee et al. in Comput Electron Agric 74:2–33, 2010). This will help in making data more transparent, reliable, decentralised and immutable. This has led to the integration of blockchain into IoT systems. The upcoming section gives a vivid picture about the basic concept and feature of blockchain technology and thereby detecting various advantages of integration of blockchain into IoT.

Internet of Things (IoT) has been continuously bringing vast technological advancements in our daily lives, thereby serving to simplify our life and making it more comfortable through its innumerable applications. IoT offers numerous benefits in the field of healthcare by reducing the cost of services and by offering care to patients that require intensive care or remote assistance. This provides numerous opportunities to enhance the quality of healthcare and reduce the cost of healthcare services. Lack of medical services assets and rising clinical costs make IoT-based innovations necessarily be customized to address the difficulties in medical services. IoT provides unprecedented advancements in the field of healthcare. This paper examines the various roles of IoT that are revolutionizing the healthcare domain by imparting extensive benefits to mankind by providing practical and affordable medical assistance. This paper presents IoT in healthcare, literature review of work carried out in this area, various challenges faced, and future scope.

Common arguments related to Delay-Tolerant Networks often focus on routing performances, energy consumption or quality of service. Despite the fact that privacy-related issues are generally considered critical, they do not arouse the same interest as the other reported aspects. For instance, when the carried information contains sensitive data, it should be protected from malicious intent. Nevertheless, most of the protocols used seem to ignore this, assuming others should take charge of the task. The aim of this document is to propose an innovative routing protocol, able to provide privacy preservation with a delivery performance comparable to the current most widely used protocols by taking inspiration from a feature borrowed from human nature, the vocal timbre.

Peer-to-peer lending is an emerging financial domain enabling people to receive instant credit facilities without much complicated procedures and intermediaries. Many financial institutions are focusing on setting up peer-to-peer lending platforms to enable hassle-free credit facilities with transparency between lenders and borrowers. The trust and transparency for hassle-free settlement and addressing novelties is a major concern in this domain which is hindering the mainstream adoption of such lending platforms. The objective of this paper is to propose a trusted and transparent distributed ledger approach using blockchain technology for setting up peer-to-peer lending platforms. The proposed approach creates cryptographically secured transactions stored over a publicly verifiable immutable ledger, which ensures credibility and auditability to investors and borrowers on every aspect of security. Proof of Immune Algorithm was proposed by leveraging the potential dendritic cell algorithm mimicking human immune system to provide consensus among peers involved in the lending process to enable trust.

Automation of vehicle classification is essential in the establishment of effective Intelligent Transportation Systems (ITS). Based on the MIOvision Traffic Camera Dataset (MIO-TCD), this paper categorizes the types of vehicles as car, bus, van, light truck, motorcycle and multi-axle truck. The classification of surveillance images is achieved using an ensemble of Deep Networks. Three networks are trained separately to make up the deep learning ensemble model with ConvNet, LeNet and EfficientNet achieving 89%, 68% and 87% classification accuracy, respectively. Results of experiments unveil that the ensemble of networks outperforms the individual networks. The ensemble of networks achieves 92.77%, which is high when compared to the performance based on genetic method in the recent literature.

One of the most important activities in a classroom is the process of marking attendance. It is a universally accepted process to measure the punctuality of a student. However, there are certain drawbacks to the existing system where a teacher has to physically mark, whether a student is present or not. The main drawbacks of taking physical attendance are time-consuming while marking the attendance and there is a possibility of proxy attendance. The existing systems like fingerprint scanning and RFID are not completely proof worthy and can be easily tampered with. Keeping this view in mind, we implement a face recognition algorithm to mark the attendance of a student. The main aim of the proposed system is to make the process efficient and save time. The proposed system recognizes the student’s face from the images stored in the database and updates it in the attendance sheet automatically. To implement the online attendance system, we use popular algorithms like Haar-Cascade and HOG algorithm which is provided by the face_recognition library. As most of the face recognition algorithms work on 2D frames, they are unable to overcome the problem of spoofing, where the person’s face gets recognized from a photo. This problem of spoofing is dealt with the help of the eye-blinking detection CNN model and trained using Keras. The proposed system uses OpenCV and Machine Learning techniques to perform the complete process.

In the era of demonetization, the banking sector has seen an exponential increase in the usage of digital payments. There has been a slew of digital payment networks proposed by both corporate and public entities. These platforms are being used by users to make payments, pay bills, and send money. The cost of Internet plans, the availability of low-cost mobile handsets, and technological savvy are just a few of the factors driving this digital revolution. Although private companies’ platforms are preferred by the bulk of people using digital platforms, public players are continually bringing novel ideas to the table, such as UPI. Another new payment platform named e-Rupi has been created and released for users by the Indian government in a similar endeavor. This platform attempts to use a voucher-based system to deliver social programs, health benefits, and a variety of other services. Hence, this paper investigates the detailed functionality of the e-Rupi platform and performs an empirical evaluation and comparative analysis of e-Rupi with other digital payment platforms.

Flying ad hoc networks (FANETs) empowered with unmanned aerial vehicles (UAVs) are a subset of mobile ad hoc networks (MANET). In a FANET, a swarm of mini-UAVs is deployed as per application scenarios to communicate critical data to ground control stations (GCSs). Owing to their distinct characteristics and unique features, FANETs pose numerous challenges making secure communication a cumbersome task. Security issues in FANETs may exist either from intrinsic design flaws or due to any extrinsic attacks performed by an attacker. Before designing the secure routing protocols, all existing security issues related to FANETs must be explored in detail. This paper explores security attacks feasible on FANETs’ routing protocols that may occur either due to network design flaws or perpetrated by a malicious attacker to gain unauthorized access to the network. The potential countermeasures against the possible routing attacks are also highlighted.

The phenomenal growth of cloud computing over the last few years resulted in the need of developing new mechanisms to make the system more secure in every possible manner. When it comes to establishing access control mechanisms in a cross-domain virtual organization-based cloud environment, the situation becomes more cumbersome due to the lack of proper standardized solutions. In this paper, an approach is being defined which combines role-based access control policy with the establishment of trust among the various nodes of different domains which are a part of the cloud environment. Several parameters are considered to calculate the trust values of different entities based on the concept of a collection of feedback from neighbor nodes. A statistical mechanism is also introduced to prevent any biased feedback. If the total trust value exceeds some pre-defined threshold, then only it is considered for the transaction; otherwise, it is not considered. The resource provider as well as the initiator can gather the feedback, and the trust is calculated from both ends in this model. We assess the trust of both user as well as the resource provider so that a proper access control mechanism can be established.

Number Plate Detection System is a surveillance system which captures vehicle images and recognizes their license number. This system helps to minimize the traffic violations. For tracking, identifying stolen vehicles and unauthorized use of vehicles, this number plate recognition technique is very helpful. A number plate detection system is used to develop a real-time application to monitor the violation of traffic rules. The first objective of the paper is to develop a system which detects vehicle’s number plate, to detect the characters on the number plate, and to show the details of the vehicle owner. To detect the number plate and recognize the characters of the number plate we used the Optical Character Recognition (OCR) technique in LabVIEW Software.

Unlike conventional Recommender Systems (RS) techniques where items are recommended to users, in Reciprocal Recommender Systems users become the items that are recommended to other users. Hence the generated recommendations should be accepted by both the service user (receiver of the recommendation) and the recommended user. In this paper, we propose SBiGRU: Siamese Bidirectional Gated Recurrent Units-based model to generate reciprocal recommendations in the online Job Recommendation by computing the semantic similarity between service users and recommended users. It can recommend reciprocal recommendations to any recruiter and job seeker even in the absence of interaction data that may not be available due to data privacy being exercised by concerned parties. The performance of SBiGRU is compared with existing state-of-the-art approaches and two Recurrent Neural Network variations viz. Bidirectional LSTM (SBiLSTM) and unidirectional LSTM (SLSTM).

The quality of images is often hampered due to the presence of noise. There are different image denoising techniques that can be used. One such technique is the use of filters. Filters are used for enhancing the appearance of images by eliminating unwanted information. We provide a detailed comparative analysis of different filters that can be used in denoising images containing various noises, in this paper. Four different noises, Speckle, Salt and Pepper, Gaussian, and Poisson, have been considered and different filters like Bilateral, Wiener, Mean, and Median have been applied to images containing each of them. The different filtered output images have been compared with the original image using their structural similarity index. Through observation and experimentation, new combinations of filters like Multiple Mean and Median-Mean have been introduced. The processing time has been calculated to decide upon the performance of different filters. A conclusion has been drawn as to which filter has to be used for denoising images containing different noises.

Currently, many applications of information search tourism are limited in the COVID-19 pandemic using a search engine. However, most application service online has not supported directly, matching end users with their preferences to find suitable tourist places. This paper has presented a proposed model using the Context Matching algorithm mostly based on the Smartphones; matching with user’s preferences and behaviors allows users to find tourism packages and regions. The experimental results show that the proposed model achieves significant improvements in matching user preferences for the domain under dynamic uncertainty. We posit that our novel approach holds the prospect of improvements in user preferences for tourism and weather in the COVID-19 Pandemic.

5G technology is a revolutionized technology of wireless communication system, and it changes the way of communication as used before. It makes it easier for the user to communicate with any devices linked to the 5G network. 5G can provide a bigger data rate, low latency, low power consumption, higher energy efficiency, enhanced quality of experience (QoE), higher throughput, etc. One of the important technologies is the massive MIMO system, which is used in the 5G network to provide an effective and high-quality signal to each and every individual user. 5G is a highly efficient network and it increases the network capacity and spectral efficiency at an extreme level. The 5G network can connect a huge number of user equipment or appliances, which transmit different signals that cause high interference in the network. So Interference becomes a major issue in this network. There are various types of interferences in this network like inter-cell and intra-cell interference, inter-symbol interference, co-channel interference, interference from various connected machines, and interference from other connected devices. For resolving this issue, different linear and non-linear equalizers are used. In this work, the performance analysis of a network with massive MIMO system is done by using zero forcing, minimum mean square error (MMSE), and maximum likelihood (ML) equalizer algorithm, where zero forcing and MMSE are linear equalizers while ML being a non-linear equalizer. The BER analysis is accomplished under the Rayleigh channel using 4-QAM modulation. The present work has been analyzed and implemented using MATLAB R2020b. The simulation results show that the BER value is lesser for higher SNR value, and the ML equalizer gives the lesser BER among these three.

We all know that the air is essential for the existence of mankind but it is getting polluted due to the sundry activities being conducted by humans. The concept of air monitoring is old but a paramount concept for all of us. Air pollution monitoring solutions were commenced long back but were still intricate in some ways. A modern technology available today is simple, expeditious, facile to implement, and gives precise data. There are different technologies that we have utilized for presenting our conception. In the present scenario, the air quality inside the house is worse compared to the air outside due to the sundry activities which we perform customarily. This is earnest issue we require to work on. Considering this situation, air monitoring is not enough. We additionally need to add purification systems that would help to live a salubrious lifestyle. The main aim and goal of this paper is to highlight some of the advanced technologies which can be used to analyze, monitor, and purify the air, how efficacious these technologies are and to study the paramount researches in this particular area. These advanced technologies are propitious in doing the analyses of air quality and find a solution to these issues that are arising due to air pollution.

A stock market, sometimes referred to as an equity market, is a gathering of buyers and sellers of stocks that represent company ownership. In this market, various investors sell and acquire shares based on stock availability. Stock trading is an important practice in the world of finance, and it is the cornerstone of many enterprises. A developing country’s rapid economic development, such as India’s, is dependent on its stock market. It is crucial in today’s economic and social environment. The stock market’s ups and downs have an impact on stakeholders’ benefits. Stock market value prediction has long captivated the interest of investors and researchers because of its complexity, inherent ambiguity, and ever-changing nature. “Stock market prediction” is a method of trying to anticipate the worth of a given “stock” in the coming days. This is performed by considering historical stock values as well as price variances throughout the previous days. Due to market volatility, forecasting stock indices is definitely tough, necessitating an accurate forecast model. Recent advancement in stock market prediction technology is machine learning, which produces forecasts based on the values of current stock market indices by training on their prior values. The term “machine learning” (ML) refers to a subdivision of “artificial intelligence” (AI) in which we train machines with data and use test data to forecast the future. This study presents an overview of deep learning techniques that are currently being used to anticipate stock market movements and predictions.

Artificial intelligence is a kind of digital representation of human intellect. Its operations are similar to those of humans in that it can make the computing machines learn, de-learn, and re-learn iteratively over and over again. The roadmap to the application of AI to Cybersecurity has been encountering a lot of issues and challenges. Digital security issues are critical in the creation of methodologies and support measures in any organization to properly prepare for countermeasures to cybersecurity threats and attacks. The role of AI in these countermeasures without a doubt is of paramount importance at present or in future. Some of the worldwide frameworks and standards being used in the context of cybersecurity across the globe may have been advocating for using AI for security. But a lot more effort is needed before the start of the process of integrating AI into these frameworks and standards and bringing the digital transformation thus caused to fruition. This paper presents an account of some of the currently used frameworks and standards for cybersecurity by various organizations around the world followed by a snapshot of using AI in cybersecurity. It also presents a discussion on the pros and cons of integrating AI into cybersecurity countermeasures.

In the Ground of Cloud Computing, one of the viable approaches to accomplish security is to utilize interruption recognition frameworks, which are programming instruments used to distinguish unusual exercises in the organization. Intrusion Detection System (IDS) has become an essential part of personal computers and data security systems. IDS normally manage a lot of information traffic, and this information may contain repetitive as well as insignificant highlights. Selecting the best quality features that represent qualitative data and excluding the redundancy from it is a key aspect of IDS. This paper proposes a new approach to Intrusion Detection Systems and creates a new fitness function for Cuckoo Search Algorithms. The proposed model has been trained with respect to KDD Cup 99 dataset.

Monitoring elderly and weak diseased people is one of the biggest issues in this modern world. Framing a technology for them is one of the wise contributions that can be done to society. More than 30% elderly people of age above 70 are falling every year due to bad health conditions. Fall identification is a significant issue in the medical care office. Elderly people are more inclined to fall than the others; accidental falls cause injuries, severe injuries and lead to death too. In our country, more than 30% are elderly people aged above 70, and they fall every year due to bad health conditions; nearly 40%–50% of elderly people fall every year most of them experiencing recurrent falls which may cause injuries, and it may lead to death too. Most of the elderly people experiencing recurrent falls which may cause injuries, to reduce the incident a system of monitoring and control is developed to detect the elderly person falls and can take immediate action, so here we considered two ways for preventing: one is smartphone-based and the other is a wearable device based recently in the fall detection wearable devices is the best choice because they are very much less in cost than the ambient-based overall features is to increase the acceptance and continue to monitor with its deep neural networks, deep learning has quickly altered the language processing domain. The LSTM is a typical recurring cell unit for deep learning models based on recurrent neural networks; here, in my paper I have proposed a new advanced version of Long Short-Term Memory (LSTM) which is Cerebral LSTM which shows better accuracy while training and testing the data and better ability to know about the time series prediction; using the RNN, the elderly person who falls is detected and with the help of the sensor the data gets collected and is allowed to training and testing validation with the MobiFall dataset; I have achieved a fall detection accuracy of about 98%.

Nowadays many new technologies have been used in various sectors like industry, medical, traffic, etc. These technologies help in controlling, managing, and maintaining the tasks in a much easier and more efficient way. The automatic license system is a real-time embedded system that automatically detects a vehicle number plate. This system has various applications like smart security systems, smart parking systems, traffic control systems, toll management systems, etc. Automatic license plate recognition (ALPR) has many complex features because of various effects. These effects include light, speed, weather, and so on. In this work, an advanced technology for the detection of number plate has been proposed. The RGB image is converted to a grayscale image. The bilateral filtering technique is used to eliminate the noise. To detect the edges canny is applied contour, masking, and segmentation are used to detect the plate. OCR is considered to read the character. The final model analysis is done to judge the accuracy. The classification system accuracy obtained is more than 90%. This ANPR system helps in reducing human intervention and provides effortless ways for license plate detection.

Nowadays, Healthcare services are dependent on Health Information Systems (HIS). The Healthcare recommendation system plays a vital role in healthcare services, work as an essential tool for decision-making tasks. Health recommendation systems improve technology accessibility and simultaneously reduce the overload of information. Although technological advancement inside the medical domain extends back years, there are still numerous difficulties to be resolved. There are various tools and recommendations for doctors in health recommendation systems (HRS). The HRS can be based on collaborative filtering, content filtering, and knowledge filtering based or may be based on hybrid filtering-based techniques. HRS is used to evaluate patient information to derive the quality of content and aid in disease diagnosis and prediction. Patients can take medicine recommendations with the help of HRS. The classification models have discussed the existing performance metrics and comparative analysis such as LSTM, Fuzzy-Logic, CNN, CNN-LSTM, etc. These classification models have improved the precised parameters as compared with the existing deep learning models. It is used to monitor the wellness and critical condition of the patients. This study is seen to be a useful starting point and the foundation for HRS literature evaluation.

The particular challenge in vehicular environments compared to other wireless networks are the rapidly changing topology and intermittent connectivity. The Named Data Networking (NDN) model has recently been proposed to offer new perspectives to meet the challenging demands of vehicular environments. The NDN exploits content name and eliminates the need to establish and maintain an end-to-end connection, which enables efficient communication in highly mobile vehicular environments. When combined with the VANET environment, the basic method of forwarding interest packets is flooding. However, this approach will lead to the broadcast storm problem which reduces the performance of VANETs’ applications. In this work, we suggest an adaptive distance-based protocol for vehicular named data networks to mitigate the interest in broadcast storm. In this distance-based approach, each vehicle dynamically computes locally an adaptive waiting timer based on the distance between the current receiver and its neighboring vehicles. The simulation result indicates that the proposed works outperform the rapid traffic information dissemination both in terms of interest satisfaction ratio and end-to-end delay.

CPU scheduling has a substantial influence on system resource usage and overall performance. Scheduling Algorithms are a technique for reducing CPU resource deprivation while simultaneously maintaining fairness among the numerous programs that utilize the resources. Round Robin is a preemptive scheduling method that significantly improves response time by restricting each operation to a certain length of time known as the Time Quantum. Various efforts have been made to calculate a time quantum value to optimize these Round Robin algorithm parameters. However, this gain in response time comes at the expense of turnaround and waiting time. In this paper, we compare the conventional Round Robin CPU scheduling algorithm to updated Round Robin algorithms such as DRRHA, as well as our suggested approaches termed MDRRHA and NDRRHA, which seek to reduce process waiting time. The Quantum value for MDRRHA and NDRRHA is derived dynamically using the arithmetic mean and the normal distribution of execution time values of tasks, respectively. The recommended solutions decrease average turnaround time and average waiting time values by up to 13%. In this research, we compare different job scheduling approaches by simulating them in a variety of test situations.

The blood–brain barrier (BBB) regulates the flow of 97.9% of the chemicals which reach the central nervous arrangement. To allow the manufacture of mind medicines for the handling of different brain illnesses, for instance, Parkinson's, Alzheimer's, and brain cancers, complexes with high penetrability be found. Several models have been created over the years to tackle this challenge, with satisfactory accurateness slashes in forecasting chemicals that cross the BBB. Nevertheless, forecasting molecules with “low” penetrability has proven to be difficult. In this research study, several machine learning classifiers such as Principal Component Analysis PCA, Neural Network SVC, and XGBoost have been compared using Molecule Net and presented in the result section. Before developing the classification model, several issues to improve the high-dimensional and unbalanced data are treated by oversampling techniques, and the high dimensionality is addressed using a nonlinear dimensionality decrease method recognized as kernel major constituent analysis has been done. A neural network with 500 epochs shows an accuracy of nearly 98% which is much better than the previous works.

The early identification and diagnosis of leukemia, i.e., the exact distinction of malignant leukocytes at the lowest possible cost in the early stages of the disease, is a key challenge in the domain of disease diagnostics. Considering the large frequency of leukemia, flow cytometry equipment is scarce, and the procedures available at laboratory diagnosis facilities are tedious, complex, and time-consuming. Inspired by the possibilities of machine learning (ML) in illness detection, the current critical search was done to examine the research attempting to find and categorize leukemia using ML algorithms. This research study provides a complete and systematic assessment of the current state of all available ML-based leukemia detection and classification algorithms that analyze PBS pictures. The accuracy rate of the ML techniques used in PBS image analysis to detect leukemia was greater than 93.5%, showing that the application of ML might lead to exceptional results in leukemia diagnosis from PBS pictures.

Early diagnosis of Covid-19 is a challenging task requiring congruous clinical medical imaging, which is a time- consuming process and suffers from accuracy problems due to variations between different laboratory results. The clinical symptoms of Covid-19 show resemblance with acute respiratory distress syndrome. The major clinical symptoms linked with this disease are fever, cough, headache, migraine, and breathlessness. Despite tremendous research going on, knowing the way of transmission and its early detection remains a mystery. There is no treatment as of now for this virus, so a lot of unprecedented containment and mitigation policies such as closure of business places, schools, and colleges, marriage gathering restrictions, transport restrictions, and social distancing are being employed. These policies are able to limit the transmission of covid-19, but are not always feasible. Steps must be taken to slow down the spread of this virus and make an early diagnosis of infection to save lives. This paper gives a clear idea about the introduction of Covid-19, its symptoms, post covid-19 symptoms, challenges posed by the virus, and proposed solutions for its early detection to slow down its rate of transmission. Methods: The proposed solution includes a clustering algorithm for massive contact tracing that helps to slow down the transmission rate, and automatic virus detection and classification network known as Convolutional Neural Networks (CNN) based upon deep domain transfer learning. Results: The pre-trained model VGG-19 is used and the hyperparameters of the model are tuned as per classification requirement by exploiting the concept of deep domain transfer learning. The model is implemented on publicly available chest radiography images and the system classifies the dataset as covid and non-covid images. The CNN achieves $$97.35\%$$ 97.35 % accuracy outperforming all the existing methods. A new concept of employing nanocoating and nano sprays is also introduced in the paper. Conclusion: To discuss various critical challenges posed by Covid-19. Addressing those issues and proposing various solutions. Proposing clustering machine learning algorithm for massive contact tracing. Developing automatic covid-19 detection and classification system based upon automatic feature detection. Providing solutions based upon nanotechnology to slow down transmission.

Learning is an ongoing process irrespective of age, gender, and geographical location of acquiring new understanding, knowledge, behaviours, skills, values, attitudes, and preferences. Formative assessment methods have emerged and evolved to integrate elements from learning, evaluation, and education models. Not only is it critical to understand a learner's skills and how to improve and enhance them, but we also need to consider where the learner is going; we need to consider navigational patterns. The extended learning and assessment system, a paradigm for doing research, captures this entire view of learning and evaluation systems. The function of computational psychometrics in facilitating the translation from raw data to concepts is central to this paradigm. In this research study, several factors are considered for psychometric analysis of different kinds of learners, and based on a motivational level, many interesting conclusions have been drawn and presented in the result section at the end of the paper.

DIP for Medicinal examination is considered as a vital topic for artificial intelligent system. We are introducing a hybrid technique combining K-means and Fuzzy C-means clustering algorithms for determining whether a brain MR Imaging scan consists tumor or not. As K-means is a hard clustering algorithm so it is used for initial segmentation via appropriate selection of the image. And after that FCM is used to provide membership to each centroid through the distance between the cluster centroid and cluster data point, prior to obtaining best result. This distance relies upon various factors, i.e., contrast, saturation, structure, brightness, and homogeneity of the image. Based upon the provided memberships by FCM technique and automated cluster selection a sharp segmented image is obtained. This modified hybrid (hard and soft clustering) approach reduces equipment and operator error. The outcome unveils that such an approach is remarkably encouraging.

5G technology is at the doorstep with the process of implementation going on as per “The European 5G Annual Journal/2021”. The 5G Network implementation is based upon the new technologies and with the new technologies even the old security issues become more and more important. With increase of data communication and with more and more devices getting hooked on to the next generation network, the vulnerability and possible options to initiate attack increases. The 5G network is expected to open up options to use this technology for mission critical applications, which makes it important to have the security systems in place. In this paper the review of various security threats those are envisioned with the implementation of 6G Network using the new technologies like Software Defined Network and Network Function Virtualization. Further, the security services may be required to be implemented in the 5G network so as to take care of the security threats that may be faced in 5G Network.

The further proliferation of age-related eye diseases, mainly age-related macular degeneration (ARMD), is increasing the load on healthcare providers. Although ARMD does not lead to complete blindness, the disease can make it difficult for people to perform daily activities such as driving, reading, writing, cooking, etc. The unavailability of any cure for ARMD, necessitates timely actions of detecting the first symptoms of eye conditions as well as following appropriate treatment options to minimize further damage. Some of the current techniques used to detect and monitor ARMD include the Amsler’s Grid, Near Vision Chart, Optical Coherence Tomography (OCT), etc. which are generally performed on paper in hospitals or clinics. This proposed solution facilitates prediction of age-related macular degeneration in patients using data collected through a Mobile application. The proposed system includes the digitization of paper-based tests as well as a novel approach for prediction of ARMD through Deep Learning. The system eliminates the need to visit a clinic and can be used by citizens from home at their discretion. The high prediction accuracy obtained while real-time testing and prediction of ARMD validates the effectiveness of the proposed approach.

Astrology is an ancient concept. Each person’s astrological chart is unique and independent, which can be influenced by different factors. In the current world, there are no standard rules or guidelines for astrological prediction. Many applications can be used to predict and analyze data, thanks to advances in artificial intelligence. These applications make use of computers to analyze unknown, large, noisy, and complex data sets and to predict and classify them. This paper aims to establish universal rules and validate astrology by using various scientific methods. This research uses the positions of stars and planets at birth to determine the profession of a person. Logistic Regression, Naive Bayes Algorithm, and Catboost Algorithm use this information to predict a person’s profession. The learning classification dataset consisted of 6248 records covering 14 different professions.

Humans can exhibit many emotions during the day, detection of these emotions can play a very important role in technology-based human–computer interaction (HCI). To achieve the detection of emotions, there are many algorithms, based on machine learning and ANN. In the proposed methodology, we are developing an algorithm, using deep learning neural network consisting of two hidden layers, which can detect various human emotions as happy, sad, angry, neutral, sad and surprise. Using the predictive analysis, the algorithm can predict these various emotions. Kaggle datasets is used for training and validating proposed algorithm. The total data set contained 35,340 data items of which 80% is used for training of the algorithm and the remaining 20% is used for testing of the algorithm. After implementing the algorithm of ReLu and SoftMax using Deep Learning neural network, results show that the model is successfully able to predict the emotion with 90% accuracy, in a variety of emotions with different types of faces. Comparison of results with other conventional approaches toward the facial expressions shows that our model does the prediction quite efficiently by extracting the data from the testing set and converting it into a JSON format which can be used as a classifier latter on without pre-processing the image data set again and again which saves both the time and computational power.

Children now start using smartphone at an early age and use different types of apps for increasing hours as they grow up. The aim of the present study was to assess the psychological impact of using smartphone on children and, in particular, determine if using smartphone for long hours leads to difficult behavior in children. We conducted the study on 130 children aged between four and ten years. We developed an app to collect usage statistics of smartphones. We asked parents to install the app on the smartphone they lend to their children and use it for at least 30 days. We also asked them to fill in the Strengths and Difficulties Questionnaire for their children. We found that 94 (72%) of the children used smartphone for one hour or more per day. Using smartphone for one hour or more per day was found to be associated with significantly (P < 0.05) higher conduct problems scale score and hyperactivity scale score among the children but had no significant effect on their emotional problems scale score and peer relationship problems scale score. Excessive use of smartphone has negative psychological effects and we recommend the use of smartphone be strictly restricted to one hour per day for children up to the age of ten years.

Assertions are frequently used to look for logically impossible scenarios, and they're an important tool for developing, testing, and maintaining software. In order to avoid a software crash or failure, software developers add assertions to their code in locations where mistakes are likely to occur. All software, even asserted software, must be updated and maintained over time. Assertions may need to be changed as well, depending on the type of change made to the upgraded program. New claims can be made in the software's upgraded version, but old claims can be kept. During regression testing of systems that utilize assertions that leverage fuzzy logic, a new approach for prioritizing test cases is described in this work for the first time. Prioritizing test cases based on fault severity, fault detection, and assessment time are the primary goals of this method. The effectiveness of a particular test case in breaching an assertion is evaluated using fuzzy logic techniques based on the history of the test cases based on the above-mentioned criteria in previous testing operations to generate the proposed solution. The proposed system was tested by simulating it in the MATLAB SIMULINK environment and measuring its performance.

Traffic Management in an optimum way seems to be an effective way to reduce traffic congestion over various intersections. The core idea behind this optimality is to provide green time for dynamic traffic flow changes in urban areas. As the vehicles are waiting in the queue during red light time, an effective control system is required to reduce the waiting time. In a fixed time/conventional traffic system green light is turned on for a fixed time in each direction. Such systems are generally pre-programmed or the fixed delay in each direction can be controlled manually and hence requires a human operator to make the desired changes, as and when required. Also, a human operator will change this for a limited number of times in a day. However, this process can be automated by using fuzzy control systems. In the fuzzy controlled traffic systems, the on time of green light is adjusted (during each transition of traffic lights) depending on the different input parameters such as Queue length, Time of the day, Arrival Rate and Waiting Time, etc. Adjustment in number of transitions indicates the flexibility/adaptive nature of fuzzy controlled system. In this paper, the novelty in the field of traffic engineering is introduced by computing the relative significance of identified parameters. Various fuzzy models with three input parameters, i.e., Arrival rate, Queue length and Waiting Time are implemented and comparative performance analysis of the seven fuzzy models hence obtained, is presented. The performance of all the implemented fuzzy models is also compared with the conventional traffic system. A traffic simulator is implemented in MATLAB to generate the real-time traffic conditions, each system is simulated and compared for all possible combinations of traffic density. Fuzzy model with two input parameters Queue length and Waiting Time outperforms the other systems and provides 23.69% average improvement in the delay observed by the vehicles waiting in the queue.

Now a day Routing Protocol in Wireless Sensor Network becomes a promising technique in the different fields of the latest computer technology. Routing in Wireless Sensor Network is a demanding task due to the different design issues of all sensor nodes. Network architecture, no of nodes, traffic of routing, the capacity of each sensor node, network consistency, and service value is the important factor for the design and analysis of Routing Protocol in Wireless Sensor Network. Additionally, internal energy, the distance between nodes, the load of sensor nodes play a significant role in the efficient routing protocol. In this paper, our intention is to analyze the research trends in different routing protocols of Wireless Sensor Networks in terms of different parameters. In order to explain the research trends on Routing Protocol in Wireless Sensor Network, different data related to this research topic are analyzed with the help of Web of Science and Scopus databases. The data analysis is performed from global perspective-taking different parameters like author, source, document, country, organization, keyword, year, and number of publication. Different types of experiments are also performed which help us to evaluate the recent research tendency in the Routing Protocol of Wireless Sensor Network. In order to do this, we have used Web of Science and Scopus databases separately for data analysis. It has been observed that there is tremendous development of research on this topic in the last few years as it has become a very popular topic day by day.

Nowadays wireless networks are being widely used due to the rise in the use of assorted electrical devices. Microstrip antenna are gaining quality for use in wireless due to their low-profile structure. A rectangular patch is designed and used for a biomedical field in this research. Rogers-RT-Duroid5880 and copper are used as a substrate and patch material. Microstrip patch antenna is intended within an operating frequency of 2.45 GHz that lies within the ISM (industrial, scientific, and medical) waveband. It gives an object’s frequency with a radiation effectiveness of 88 percentage. Performance analysis of intended antenna in terms of bandwidth, return loss, gain, radiation pattern, directivity, total efficiency, and power analyzed by using advanced design system software (ADS).

Phonocardiogram (PCG) plays an important role in the initial diagnostic screenings of patients to assess the presence of cardio-vascular abnormalities. It is also used to complement the ECG-based cardiac diagnosis for detecting cardio-vascular abnormalities. This task has been proposed to classify the heart sounds by performing deep learning technique known as CNN (Convolutional neural network). Two types of datasets were collected from the clinical environment and used as an input to this project. The experimental results show CNN provides the better results in the detection of abnormal heart sounds with good accuracy.

With the unexpected rise of the COVID-19 pandemic, a vast majority of organizations and institutions have shifted to online modes of communication, especially educational institutions. Many hurdles popped up in due course of time, preventing the successful conduction of online examinations. Today, there is an application for each and everything and the number of users is steeply escalating. With such high rising demands and new necessities, it has become quite important to make sure the applications are easily available anytime and anywhere. Ensuring minimum downtime, frequent updates without any disruptions, and making it easily available for both developers and consumers are must-haves now. Satisfying these requirements over a monolithic architecture is not assured, hence, the microservices architecture is adopted, thus opening the door for new possibilities. This being the foundation of this project, an amalgam of how the web application we used is divided into microservices, containerization of each of them followed by pod deployment and finally using Kubernetes for service exposing and orchestration is presented in this paper. In short, an online exam monitoring system that can identify any abnormal behavior by test takers while being capable of tackling varying traffic loads is fault-tolerant and handles disaster recovery.

Genetic Algorithm (GA) is a very efficient and stochastic algorithm inspired by biological evolution process. It has been extensively applied to a variety of practical applications. Convolutional neural network (CNN) is the most popular deep learning technique and has yielded the most promising results in every field. In this research paper, the restricted characteristics of GA have been hybridized with the basic features of CNN to obtain a better and improved version called Restricted Genetic Algorithm based on Convolutional Neural Network (RGA-CNN). The proposed approach is tested on two different stress data sets collected from different colleges and universities. The experiential results are compared to the state-of-the-art methods such as convolutional neural network, long short-term memory (LSTM), and recurrent neural network (RNN). The results reveal that the proposed approach outperforms CNN, LSTM, and RNN in the diagnosis of stress.

The world is moving at a fast pace and people do not usually find the time to read a newspaper or go through the news every day. A survey shows that a huge amount of time is spent commuting to workplaces and educational institutions. To convert that time into a productive one, we have developed an application that can work hands-free. This is a news application that works on web browsers both on mobile as well as desktop computers. The aim is to create an app that fetches news across various sources all around the world and displays it in an organized manner with a clean User Interface along with some hands free voice assistant features.

Several bridges and roads have deteriorated and become impassable in Guinea in recent decades. Because? The degradation of their supports (soils) and the crucial lack of basic geoscientific studies of their direct environments. We have set ourselves the objective of analyzing this problem in Middle Guinea and providing concrete solutions. The results of our social and state surveys (climatic, hydrological and geological services) show seasonal damage due to downpours. Our geoscientific studies reveal pedogenetic products (chemical and mineralogical) forming soil horizons (pasty and slippery on the surface), very sensitive to hydrolysis which modifies their states of stability (formation) and produce results which prevent the masters of ‘works on the realities that await them and guide them to a safe approach. Our geotechnical outlook provides concrete/direct measures combating this scourge on the Labé-Mali axis; at the Linsan and Mamou-Linsan deviations, thus facilitating socio-professional journeys which have become difficult.

The coronavirus was declared a pandemic by the World Health Organization, and a vaccine for it was developed and is currently being used to vaccinate people all over the world. From the beginning of the COVID-19 vaccination program, many people have refused to accept the vaccine due to widespread misconceptions and propaganda concerning the COVID-19 virus itself, whether it is real or not, and the vaccination program as well. All these misconceptions and propaganda have been spreading through word of mouth (verbal) discussion among citizens. In this study, a hybrid machine learning approach was proposed with the help of natural language processing to build a sentiment classification and prediction system using the data collected from the public regarding their opinions on the COVID-19 vaccination program. The data was collected through the Google Form. The evaluation metrics used to measure the effectiveness of the proposed work were accuracy, precision, recall, and f-measure. Various machine learning algorithms were used for the implementation. Two algorithms, namely Support Vector Machine and Bagging Classifier, outperformed the remaining algorithms with the same accuracy and precision scores of 75%, respectively. The two algorithms were considered for the voting, which served as the final hybrid machine learning model for the sentiment classification and prediction tasks. The voting classifier works by predicting an output voting class based on the highest likelihood of the combined models, which were Bagging Classifier and SVM. Using the voting classifier, accuracy, and precision scores of 75% and 75% were also obtained.

Automatic text summarization is a data-driven technique for coping with today’s huge amounts of textual material. The purpose of this study is to look into a novel way for generating abstractive summaries of huge documents. The lack of precision gained by very intricate models seeking to provide abstractive summaries necessitates the development of a new strategy. In this paper, we looked into combining extractive and abstractive summaries in a two-step process to get more accurate final summaries. The purpose of this research is to make the process of writing a summary as simple as feasible while still maintaining accuracy.

In today’s era of development we always come across situations where we are actually running our application and it suddenly crashes. At that point K8trics come into picture. K8trics (Ketrics) will be a Kubernetes (K8s) native metrics aggregator which will leverage Linux Kernel’s eBPF capabilities to efficiently capture the data from the kernel space. Collect environment and service aware metrics from a distributed system. Network Metrics like SYN timeouts, TCP retransmissions, DNS misses, Req/sec, and request latencies (p. 50, 75, 90, 95, 99, 99.9). Application Level Metrics like dynamic logging, USDT, resource usage, and CPU profiling. Service aware policy enforcement Network policies: K8trics in conjunction with Hyperion can support extremely complex network policies but the goal would be to be able to present a POC firewall. Application Level policies: K8trics in conjunction with Hyperion can support extremely complex application level policies but the goal would be to be able to present a POC socket blocker.

Communication related to data transmission. These data transformation between devices involves different approaches like wireless, wired. The modern-day evolution of the communication between each person reaches its soaring heights. Network efficiency, flexibility, and speed are all enhanced by wireless technology. The tremendous usage of mobile handheld wireless devices simplifies the human needs and leads to the innovation of 5G which enables a new kind of network designed for universal connectivity of everyone and everything, including machines, objects, goods, and devices. As a result of fifth generation the dream of connective machines, devices and objects has become true. Here a 28 GHz of resonant frequency is used to create rectangular shaped patch antenna. The designed antenna is verified at the simulation level in ADS (Advanced Design System) 2019 simulation tool. Because of its high frequency, we are using Rogers RT Duroid 5880 as the substrate which is selected as the dielectric material for the patch antenna and it’s dielectric constant is 2.2 $$\varepsilon r$$ ε r . The used dielectric material height is 0.5 mm.

Advanced speed breaker is a system that helps to promote road education, contributes to respecting speed limits, benefiting road and drive safety, with the aim of preventing accidents and raising awareness among drivers of respecting speed limits, also compiling statistics and making possible measures impact and benefits. Advanced Speed Breaker could be a business safety system where dashing vehicles spark the speed swell and rises the speed bumps on top of the paved surface and giving the physical remainder to motorist to decelerate down the vehicle. If the speed of the on-going vehicles is inside the predetermine limit then the speed bumps keep flat on paved surface and vehicles passes over it well. This configuration detects the vehicles that circulate respecting the speed limit allowed in the area and lowers the device to ground level for them, but leaving it elevated for those who do not respect predetermine limit. Additional modifications can also be made to make emergency vehicles accessible.

Internet of Things (IoT) has emerged as a powerful communication and networking system for smart and automation processing. With the increasing usage of the Internet of Things in numerous critical activities, it is essential to ensure that the communication among these devices is safe and secure. The biggest threat to safe and secure communication is from cyberattacks. Cyberattacks have evolved and become more complex, henceforth posing increased challenges to the data integrity, communication security, and confidentiality of the data. With its success in detecting security vulnerabilities in a communication network, intrusion detection systems are best integrated for securing IoT-based devices. But the integration of an intrusion detection system in an IoT-based network is a challenging task. This paper investigates the state of the art of IoT and intrusion detection system, the technology in use, and the technology challenges by reviewing notable existing works. A systematic literature review of 25 sources comprising 22 research papers and articles covering the threat models, intrusion detection system key challenges in IoT, Proposed models, and implementation of models, reviews, and evaluations are reviewed. The findings explore the needs and the best ways of integrating artificial intelligence-based intrusion detection systems in IoT networks for ensuring security and safety of communication.

Today there are a wide variety of methods of teaching, but choosing the most efficient one for a particular group of students is an effortful task. The proposed system estimates the degree of student engagement by analyzing various facial features extracted from the video dataset, time of the day, and relative intelligence determined by a questionnaire given to each student. Our model encompasses the usage of both facial features and psychological factors to achieve the result. We classified the degree of student engagement into four levels—not engaged, partially engaged, engaged, and fully engaged, ranging from low to high, respectively. The performance of the model is better with Random Forest Classifier with an accuracy of 76%. In this paper, we majorly focus on improving the quality of education by enhancing student engagement throughout the course. Based on the results, the faculty can update their teaching methodology.

According to a current evolving real paradigm in neuroscience innovation, humans may use brain impulses to interact with, effect, or transform their environment. People may use the growing brain–computer interface (BCI) technology to control, communicate, and monitor their objects by utilizing or interacting with assistive devices. BCI technology will surely rely on improved signal capture and clear validation of actual research and delivery models in the future, which will be connected to the dependability issue. For the BCI system to work better, an appropriate signal processing approach must be used that makes it easier to collect physiological data and a higher classifier suitable for the specificity of the system. We offer a concise overview of several signal processing strategies for improving BCI focus. A supervised classification method is utilized to improve the support vector machine in order to recognize and categorize EEG data. The EEG data’s temporal domain properties are retrieved and input as feature vectors into the SVM, which are subsequently utilized for classification and identification. The algorithm’s exploration capabilities and convergence time have both improved, as shown by statistical analysis of the data. The SVM parameters are then optimized using this information. The approaches for feature extraction and classification are covered in this study. This outcome represents a 2–5% improvement over the previous technique. EEG (Electroencephalography) is a complex bio-electrical signal. Researchers may uncover useful physiological information if they do a thorough examination of this data. The challenges that BCI data processing designers face are discussed here, along with some drawings of possible existing and prospective solutions. We create ways to extract and identify specific traits from the standpoint of BCI systems, and we conclude with a thorough conclusion and interpretation.

Rainfall prediction is the highest research priority in flood-prone areas across the world. This work assesses the abilities of the Decision Tree (DT), Distributed Decision Tree (DDT), Naïve Bayes (NB), Random Forest (RF), Support Vector Machine (SVM), K Nearest Neighbour (KNN), and Fuzzy Logic Decision Tree (FDTs) machine learning algorithms for the rainfall prediction across the Kashmir province of the Union Territory of Jammu & Kashmir. On application of Machine learning algorithms on geographical datasets gave performance accuracy varying from (78.61–81.53)%. Further again machine learning algorithms were reapplied on the dataset without season variable yet again performance ranged in between (77.5–81)%. Vigorous analysis has established that these machine learning models are robust and our study has established that the dataset reaches performance stagnation and thus resulting in performance capping. The stagnation is irrespective of the choice of algorithm and the performance shall not improvise beyond a specific value irrespective of the choice of the machine learning algorithm.

In recent years, there has been an urge for development in network technologies and that has contributed to an increase in cyber-attacks that are threats and challenges to the protection of network resources. In contribution to the protection of the network, artificial intelligence has been shown to be an efficient and better technique used in recent years for better detection of network attacks. In the paper, we propose an overview of deep learning techniques which are applied in intrusion detection systems. A summary of different deep learning techniques and their applications in intrusion detection systems are proposed with the various problems encountered by network security. We also give a brief summary of the benchmark datasets used in the deep learning techniques and provide a comparison of the performance of the different techniques. Finally, we propose suggestions to improve the performance of those deep learning in the attack detection.

Analyzing the multiple relevant documents returned in reply to an end-user request by an information retrieval system is challenging. It is very time-consuming and less efficient to find analogous web pages without applying the clustering. Clustering of web pages arranges a large number of web documents into relevant small clustered groups. In this paper, a novel similitude degree computation technique is proposed to provide the web documents related to the context in which multiple related web documents are the members of the same cluster. The clustering module results in web documents’ arrangement with their associated topic and corresponding computed similitude or similarity score. This provides the user clusters containing equivalent web documents related to the issue of desire. This context-based grouping of web documents reduces the time taken for searching relevant data and improves the results in response to a user request. Moreover, the comparison and analysis of the proposed technique are done with different existing similarity measures on the basis of performance metrics purity and entropy. It has shown the proposed scheme provides better results to the user.

Control and monitoring of essential infrastructures like nuclear plants, power generation and distribution plants, oil and gas, and many more facilities are handled by industrial control systems (ICS). The real-time response, high-performance computing compatibility as well as security. There are a number of protocols in place to assure the safety and security of the operations in question. ICS protocols phasing so many problems due to their high demand in real time. ICS has adopted Internet-based technologies and most communication protocols have been rewritten to work over IP as a result of the increased access to the Internet world for business purposes. In contrast to typical IT systems, the ICS components and communication protocols were vulnerable to cyber-attacks because of their openness. In order to make it easier to analyze the danger of ICS protocol cyber-attacks. We propose the taxonomy model to identify the attacks on different protocols of ICS based on the security pillar, attack type, and different protocols used in the devices. This paper help researchers and industrialist understand the attacks and issues on the different protocol.

Game is one of the most entertaining shows for today’s all generation peoples, particularly Football in most part of countries of the world. Football as a sport is only growing more and more popular every day. It is currently the world’s most-watched sport and has the highest viewership audience. As a result, a whole industry has arisen around this sport with one important part of it being FIFA. The amount of budget allocated and the number of persons involved in a Football game directly or indirectly can affect the financial budget of a person to a federation's finance. In such cases, player selection for a finalist from the federation is the most crucial task. Every year different approaches were investigated for player selections, but none of them was regarded as the best approach for team selection. Thus, there is a need for a standard approach for finding out the perfect players for their teams with the exact qualities that they demand. In response, we have developed a machine learning model that predicts players who could replace a current existing player in a team. Along with that, we have also incorporated Data Analytics that helps us decide which factors would be more important than others. The proposed prediction model is implemented and the results of our machine learning (SAGA-ML) tool are applied to Electronics Arts’ FIFA Soccer game.