ICT Analysis and Applications

Satellite navigation system has come a way long from its initial stages when it was used for military applications to now in mobile devices worldwide. With the advancement in the satellite navigation system, various new navigation satellite constellations have been set up in space. Various data processing and analyzing tools have been developed for the systems. The number of this software, both online and offline, has been increasing due to which it has become essential to have a detailed comparative study on this software to design more efficient ones in the future. To this end, this paper surveys different software for global navigation satellite systems. The selection of the software for the survey is based on their attractiveness among scientists, results published in literature, and noteworthy characteristics and features. The survey work aims to assist scientists, researchers, and software developers in selection of an apt software for their work based on system requirements, supporting constellation, supported data format, price, size, strengths, and weaknesses. Software developers can further identify limitations of the existing software and overcome them.

The digital structuring of a language depends on how their parts of speech are classified to make the language the most compatible for the user. The parts of speech (POS) of a language is defined as the word classifiers which classify a word more precisely for using it in the sentence properly for expressing the emotions or feelings of human beings through natural languages. The natural language processing in the research arena already classifies the exhaustive forms of POS of many languages in the world. Among all such languages, the POS of SANSKRIT language is also introduced by many eminent scholars. Here, an attempt is made to give the new set of POS of the SANSKRIT language with their proper definitions and explanations with examples and they are termed as Universal POS (UPOS). The set so defined that universally everybody can understand, and the further use of these tags can be made available to the researchers of the world. The new set is highly expected to be the most exhaustive form.

One disadvantage of computer-assisted detection systems is the massive quantity of data needed to train them, which is costly in the medical industry. A big training dataset is critical in deep learning since it enhances training accuracy. Even with a big amount of data, a weak algorithm can be more accurate than a strong algorithm with a little amount of data. When verified on a different unobserved dataset, data augmentation generates new data which is used to train the model and enhances performance. We presented a thorough evaluation of the literature in which data augmentation was employed to train a learning model using lung CT images. Basic and deep learning data augmentation techniques were used to categorize the articles. The term “data augmentation” states a group of approaches for increasing the volume and quality of training datasets. Geometric transformations, kernel filters, color space augmentations, random erasing, mixing pictures, adversarial training, feature space dataset augmentation, meta-learning, and GAN-based networks are among the image augmentation processes explored in this paper. Students will learn how to employ data augmentation to improve model productivity and expand small datasets in order to take advantages of big data.

Linear regression is a well-known method used in statistic to determine correlation between two or more variables, in machine learning for data classification and for prediction. This paper deals with data classification using a composition of two techniques of classification such as mesh generation, followed by the Hough transform method, in order to define a linear regression on the dataset. The purpose is to use analytical straight lines as a regression technique. The proposed method is a technique of data reduction through a mesh generation, which creates a virtual grid where each continuous point is localized in a cell corresponding to a pixel in the virtual grid. The standard Hough transform method establishes a relation between an image space and a parameter space through the definition of a sine function. The values of variables are represented by the coordinates of continuous points in the image space. The standard Hough transform is applied to each cell. The coordinates parameters with the high number of votes superior to a threshold that appears in the parameter space or the accumulator correspond to the parameters of analytical straight lines. Straight line recognition achieves linear regression and establishes correlation between the initial variables. Our analysis presents difference between classical linear regression and the proposed alternative which gives more visibility between data relation and accept a level of error in the dataset.

Internet of Things has revolutionized the entire world these days. With the advent of the word smart in all the fields, for example, the smart city, huge amount of data gets accumulated every single day. For any application, processing this huge amount of data is very important. Prior to processing, acquisition of this bulk amount of data is very important. In the existing research, there are different ways to collect these data. In this paper, data acquisition using different modern approaches is outlined. These approaches encompass data acquisition using hierarchical deep reinforcement learning, data acquisition using energy efficient UAV, data acquisition using swarm intelligence. This paper outlines these approaches, brings out the assets and liabilities of each one of these approaches for smart com.

An important task in education field is discovering student behavioural patterns to take timely action to improve student activities or grades. Sometime students may fell into depression due to misunderstanding of subjects or due to low grade which leads into abnormal behaviour, and by identifying such abnormal behaviour, institutions can take necessary steps to improve student’s condition. For this research, questionnaire method is used which includes collecting student data through survey and analyse students’ behavioural patterns. However, results by this method are not effective or accurate as this method largely relies on feedback data. So to solve this problem, an unsupervised clustering approach can be used. This produces relatively accurate results. The proposed framework integrates two unsupervised clustering approaches, i.e. density-based spatial clustering of applications with noise (DBSCAN) and k-means. The students data is collected from Kaggle data sets. The proposed framework extracts necessary behaviour features by statistics and entropy to find both anomalous behavioural patterns and main stream patterns. To predict whether the student is low active or high active or medium active, we can use supervised techniques as unsupervised clustering approaches are meant to form clusters. These findings can help students to improve their grades and personality and organization can also take appropriate steps to help students by providing better services and administrations such as psychological consultations and academic advices.

Integrated IT service management (ITSM) follows ITIL service operations process to track service request, incident, problem, or change in existing application or infrastructure stack. These tickets are raised by end-user, developer, and tester or through proactive monitoring sources to report a problem or request new feature development. This work contributes to AIOps offering for site reliability engineer (SRE) team who needs to continuously improve the application availability, security, and performance by analyzing re-occurring issues, frequent changes, performance metrics, etc. This solution also has feedback mechanism which continuously trains the model on result relevance. This research work will help the SRE to identify re-occurring issues irrespective of language, accelerates root cause investigation using historical resolution details and permanently fix the problem hampering the application performance. It is noticed that 80% of the resolution time is spent in diagnosing and detecting the issue. With knowledge of historical similar issues and resolution information, the time to detect and troubleshoot reduces drastically from 80 to 20%.

Energy Management Solution (EMS) sounds familiar in recent days for its advantages and solution to monitor and analyse energy consumption. EMS plays a vital role in the energy sector and in the field with bottom-line business priority. This paper proposes an EMS solution to the smart microgrid through an Internet of Things (IoT)-based unified framework. The IoT-based unified framework provides energy efficient optimization and scheduling to the smart microgrid. This work also discusses the energy harvest and energy trading in the rural area where the smart microgrid has been installed. The advantages of the proposed IoT framework are it provides adequate control on the renewable energy resources and optimize the load scheduling based on the energy harvest. To prove the EMS using the developed framework, a simulation analysis has been carried out. The obtained simulation results show the essential requirement of energy management solutions and the optimal performance of the proposed IoT framework.

This work is to highlight the various methods and approaches to conduct anomaly detection in application log messages in a typical software defined networks. Work also compares the precision, accuracy and sensitivity of various supervised and unsupervised methods of anomaly detection. Typically, in large scale software defined networks, anomaly detection is an important entity in managing it. Log records are widely used for anomaly detection to record system runtime information. As part of this work, guidelines for adopting this work and scope for future works are proposed. The review can assist in anomaly detection method and avoid doing any redundant works. The steps involved in the process are: log collection, log parsing, feature extraction, and anomaly detection. More specifically, reviewed anomaly detection methods include supervised methods (Logistic Regression, Decision Tree, and Support Vector Machines) and unsupervised methods (Log Clustering, Principle Component Analysis, and Invariant Mining). They are being discussed in detail along with comparison results.

Water safety is a serious problem, particularly in highly populated cities, because pathogenic microorganisms in drinking water can trigger deadly outbreaks. Bacterial contamination in drinking water, like Shigella flexneri, Vibrio cholera, Salmonella enterica, E. coli, and others, is a major source of pathogenic pollutants. The current study proposes a photonic crystal (Ph.C.) micro square ring resonator (MSRR) sensor that is small in size and has a fast output response time to detect the signature of single waterborne bacteria. The bacterial contamination in water varies the refractive index and in turn resonance wavelength shift at the throughput port is observed. The transmission spectrum is determined by using FDTD simulation of Ansys Lumerical software for the suggested structure and the impurity is determined by examining the resonance peak wavelength shift. The proposed structure is ultra-sensitive and compact in size having sensitivity of almost 946 nm/RIU and the structure would be useful for a variety of sensing purposes.

Wireless sensor network is now being increasingly popular in today’s world and is getting used in many kinds of fields and applications where different protocols are being used for the better communication. There is also a great requirement for the improvement of different parameters related to the different issues in this domain. This paper specifically discuss the performance related to increasing the energy efficiency of the hybrid protocol. In new hybrid protocol design which will increases life and energy efficacy of the network using the specifications of the various network QoS tools more particularly system concerns with network simulator NS2 and different parameters and issues regarding the same.

The proposed AI-based Performance Analysis and Navigation System (AI-PANS) aims at the development of self-learning app for the students of Grade 1–10 students studying in the Maharashtra (India) State Board and also for the teachers. The proposed AI-PANS is designed for: – Database Generation, to ease the process of making new questions. – Recommend question paper based on the student’s current and past performance in every iteration. Modify/redefine the Degree Of Difficulty (DOD) based on the time taken and correctness of the attempted question by a large set of students. – Navigate the student across topics based on his/her performance and in turn help him/her in clearing the concepts, improve the problem-solving skills, build confidence, retain interest, and prepare him/her for competitive exams. – Hand-holding in the form of curated solutions in the form of text, images, audio, etc., and summary of each topic. In the proposed system, we also proposed a novel mechanism to redefine the DOD assigned to every question in the beginning.

The distributed denial-of-service (DDOS) exploit is one of the most catastrophic assaults on the Internet, disrupting the performance of critical administrations offered by numerous organizations. These attacks have become increasingly complicated, and their number has been steadily increasing, making it harder to detect and respond to such assaults As a result, a sharp security system (IDS) is necessary to detect and control any unexpected system traffic behavior. In a DDOS Assaults, the intruder delivers a stream of packets to the server while exploiting known or unknown flaws and vulnerabilities.

Upcoming wireless networks like 5G, Wi-Fi 6, and beyond are projected to be exceedingly complicated and dynamic. The rise of ultra-dense complex network deployment, high data rates, with new applications may necessitate a new wireless radio technology paradigm may offer several critical challenges for network administration, operations, and planning, including troubleshooting. Similarly, the generation as well as consumption of wireless data is shifting from persons to machine-oriented communications, considering future wireless network operations even more complex. Like a result, new approaches for deploying dispersed computation means with greater context awareness will become more important just for reduce the complication of future wireless network. Studies indicated that has been focused on wireless broadband connection of the 5th generation, called “5G,” which is now implemented through mobile network operators. Unexpectedly, ‘Wi-Fi 6’, is the newest IEEE 802.1ax model inside group based on wireless local area network techniques including characteristics aimed for private, and edge-networks, has received far less attention. This paper examines the potential for cellular along with Wi-Fi networks for providing speedy wireless Internet connection.

In the current world of automation, everything is getting atomized as reducing manual labor is the key to efficiency. Here, we focus on offline handwriting recognition. Handwriting recognition is the process of extracting text from handwritten scripts, this is also known as offline handwriting recognition. The purpose here is to attempt to improve the accuracy and efficiency of the system using neural networks and datasets used for training the model, as well as detecting and identifying the characters and exporting them in a text format. The proposed system can be used to recognize handwritten characters and convert them into the text from the scanned image of a page.

A graph G(V, E) is called total magic cordial graph if it owns up the labeling called total magic cordial labeling which is defined as, a function $$f:V\left( G \right) \cup E\left( G \right) \to \left\{ {0,1} \right\}$$ such that $$f\left( u \right) + f\left( v \right) + f\left( {uv} \right) \equiv C\left( {\bmod 2 } \right)$$ for all arcs uv $$\in E\left( G \right)$$ provided the condition $$\left| {f_{0} - f_{1} } \right| \le 1$$ is hold, where $$f_{0} = u_{0} + e_{0}$$ and $$f_{1} = u_{1} + e_{1}$$ where $$u_{i}$$ , $$e_{i}$$ , $$i \in \left\{ {0,1} \right\}$$ denote the nodes and arcs, respectively. This research article found that octopus graph, vanessa graph, lilly graph, and lotus graph admits the above mentioned labeling for all values of n.

Missing data cause many challenges for imputing missing value in real-world datasets. Much research has been done on these challenges, but most existing research focuses on the classification task. Only a few methods can handle the large datasets for imputation of missing value. This paper proposes a new hybrid approach to impute the missing value. This approach is based on a Weighted k-nearest neighbor (WKNN) and genetic programming algorithm. This approach aims to enhance the accuracy of the imputation of missing value in symbolic regression. This paper has used different datasets with a different missing ratio of data and applied the imputation model to the datasets. This approach makes exact imputation compared to other methods like Decision Tree, Genetic programming imputation, Bayesian Regression, logistic regression, WKNN, Multilayer Perceptron, Random Forest, and Support Vector Regression.

In a densely populated country like Bangladesh, the groundwater levels have dramatically decreased in the past few decades while arsenic concentrations in them are increasing. In many cases, a significant amount of water is wasted due to overflow, resulting in a loss of valuable water resources as well as money. Water conservation has become more important than ever, while alternative water sources must be considered. To solve this, it is best to monitor the level of water under consideration by using IoT alongside collecting rainwater. This can help reduce the stress on groundwater in a country like Bangladesh, where we receive an annual average of 2200 mm of rainfall every year. The main obstacle, however, is that due to heavy air pollution, an undesirably high concentration of gases such as sulfur dioxide and nitrogen dioxide mixes with rainwater and causes acid rain. This research aims to achieve sustainable development by evaluating rainwater harvesting and ensuring the collection is safe for use.

Identification of drowsiness while driving requires continuous driver monitoring, and wake-up alert modelling, which assists in reducing on-road accidents. To perform this task, multimodal analysis including, eye tracking, yawn tracking, oxygen level monitoring, road parameters, etc. are evaluated, and machine learning models are applied to them for classification of drowsiness conditions. But most of these models have limited accuracy and precision performance when applied to real-time conditions. This limitation arises, because the models are trained on clinical datasets, which lack in feature density under real-time environments. To overcome this limitation, a novel fusion-based model for identification of drowsy drivers via cascaded deep transfer learning is proposed in this text. The proposed model initially uses standard datasets for eye tracking, yawn tracking, electroencephalogram (EEG), and driving behaviour to train an ensemble learning model, which is fine tuned via on-road conditions. Data collected from on-road conditions is processed via a correlation engine which assists in similarity identification and transfer learning processes. The proposed model uses mel frequency cepstral coefficients (MFCC) with i-Vector features and classifies them via augmentation of k-nearest neighbours (kNN), random forest (RF), linear support vector machine (LSVM), logistic regression (LR), multilayer perceptron (MLP), and customized 1D convolutional neural network (CNN) classifiers. Results of the model were evaluated in terms of accuracy, precision, recall, f-Measure, and response time metrics, and were compared with various state-of-the-art methods. It was observed that the proposed model showcased 8.3% better accuracy, 6.5% better precision, 3.4% better recall, and 14.5% faster response than existing models. Due to this performance enhancement, the proposed model is deployable for a wide variety of real-time application scenarios.

Heart disease refers to the condition when the heart is not capable to push required amount of blood to the entire body. Heart disease (HD) is the prevailing reason behind deaths among the world-wide population. Early prediction of heart diseases can save lives. Predicting cardiovascular or heart disease in advance, a person can be warned beforehand, and the death can be prevented in turn. Machine learning (ML) has made a huge contribution to classify the population with heart disease from the healthy population. This paper proposes three heart disease prediction (HDP) models namely LOFS-ANN, LOFS-SVM, and LOFS-DT utilizing lion optimization-based feature selection (LOFS) method and three ML-based classifiers. The datasets used are from UCI repository. The comparative analysis reflects that the model LOFS-ANN performs best among all three models, with the values of 97.1% and 90.5% for AUC measure and accuracy measure, respectively. It can be concluded that the LOFS-ANN has a significant potential to predict heart disease after drawing its statistical comparison with the competing models.

In this era of deep learning and artificial intelligence, where a large scale high-quality dataset is a must, we present a high resolution (1024*1024) synthetically generated latent fingerprint RUVIS dataset with pore level details. It will assist the research community to overcome the privacy issue concerns which hinders the large scale manual collection of fingerprint datasets. For this, the training dataset has been collected using RUVIS equipment which works on the principle of reflection of ultraviolet light, whereas synthetic prints are generated using StyleGan2-Ada. Our experimentation suggests that the synthetically generated latent fingerprints are not only diverse and unique but also highlights pore level details. Further, the experimental analysis has been done using structural similarity index (SSIM) and multi-scale structural similarity (MS-SSIM).

Blockchain technology has emerged as a new security technology that offers a variety of benefits to numerous businesses. Blockchain is working on the concept of distributed ledger technology. Multiple industries are adopting blockchain technology due to its different features like distributed networks, consensus mechanisms, and transparency. Blockchain technology gives a great impact on the business environment. Blockchain technology does not involve third-party interference and provides security and transparency to data. The data inside the blockchain is immutable, and no one will be modified it or delete it. Smart contracts and consensus algorithms are the major components of blockchain technology. A smart contract is a sort of computer software that establishes agreements between the nodes of a blockchain network’s participants. Few major consensus algorithms are PoW, PoS, PBFT, DPoS, and tendermint. This paper evaluates the different consensus algorithms and principles behind using it. Paper demonstrates the steps to be followed for the implementation of consensus algorithms.

As the frequency of security breaches continues to rise, cybersecurity remains a critical concern for every industry in the online. Thousands of zeroday attacks are known to emerge on a regular basis as a consequence of the integration of multiple protocols, primarily from the Internet of Things (IoT). The majority of such attacks are minor variants on previous research findings intrusions. This suggests that even sophisticated techniques like typical machine learning (ML) algorithms have difficulties spotting these tiny kinds of attacks over time. These attacks are called as DDoS attack; they are used to prevent clients from accessing a server or website. DDoS attacks have been employed by cybercriminals to bring down targeted servers and breach venture networks with the ability to overwhelm results. Because of the growing volume and complexity of DDoS attacks, many organizations are having difficulty handling them. Smart gadgets and IoT are particularly vulnerable to a wide range of DDoS hits due to resource constraints such as limited memory and processing capacity, thus cybercriminals are aware of these current technologies and their flaws. Because of an attack on their internet service providers in 2016, many firms, including Netflix, CNN, and Twitter, were forced to go down for nine hours. This technological failure resulted in several issues, including financial losses, productivity losses, brand damage, insurance rating drops, unstable client-provider relationships, and IT budget overruns. We need to construct an IDS system to expose and prevent DDoS attacks to secure data processing, information technology, and commercial components. The cost of cybersecurity will be greatly lowered if security teams use current and new technology like ML, automation, and artificial intelligence (AI). This study will examine the detection performance of DDoS attacks using several ML, DL techniques and also categorize it into cloud and fog ecosystems.

The pandemic has hit many businesses from March 2020.This effect remains over a year and may continue further depending on the efficacy of the vaccine and the variants produced by the virus. The construction industry which is completely dependent on labour (unskilled, semi-skilled and skilled workers) has had an impact. We studied this impact and analysed the main causes of the damage. This work is on two incidences, the first and the second lockdown. Areas selected in Mumbai city. The sector has been divided in three verticals, Real Estate (Organized), Real Estate (Un-Organized) and the Supplier Segment. Our study reveals an impact due to three main reasons, labour movement, material cost and lock-down restrictions. It is clearly visible that the movement of labour is better in the organized sector than the un-organized and supplier segment. This leads to the fact that we as a country, are moving towards industrialization of the construction segment, where work force is preferring facilities and benefits over pay. A positive change displayed in our study and the process towards westernization is certainly a positive way forward.

Over the years, the issue of linguistic variation has posed a significant barrier to successful information collaboration. In modern times, learning multiple languages can be both time consuming and a hectic process because of the language difference, and it can also be an expensive expenditure if a tutor is appointed for teaching the same. Language interpreters must be able to communicate in and understand both languages, because the traditional approach to handling the problem of translation is neither productive nor advantageous. TranslateIT is a mobile application that can be used to translate among 11 native languages that includes English, Afrikaans, Arabic, Belarusian, Bulgarian, Bengali, Catalan, Czech, Hindi, Urdu, and Welsh. Therefore, we have developed an android-based mobile application called TranslateIT which is a multi-language translator which has been made using firebase ML in Android Studio using Java. Because of this application, tourists will be able to communicate more effectively with locals and gain access to more relevant information, which will help them learn more about their surroundings.

With the rapid development in the computer science and technology domain, the eruption of available data is observed in this decade. Online platforms are becoming more and more capable day by day and therefore can capture thousands of customer reviews and comments on a single post. Dealing with this gigantic amount of high-dimensional text data leads to several problems for the post owner and data analysts. Furthermore, a significant percentage of this high-dimensional text data is not important and can be proficiently concentrated to lower dimensions by using several advanced dimensionality reduction methods. Topic modeling methods are used to summarize text data efficiently and are a good way of analyzing a huge amount of text data. In the recent years, many topic modeling approaches are introduced and are used to gain fruitful insights from the considered dataset. This paper aims to suggest the most effective known model by analytically comparing several existing topic models on the taken dataset. Also, some modifications to the existing algorithms are suggested to generate more understandable and accurate results.

This paper presents a brief survey on different modulation schemes of a visual light communication (VLC) system, based on which the performance of different modulation schemes has been compared.

In this paper, we have suggested a method for simultaneous quantification using high-performance liquid chromatography for Nebivolol and Cilnidipine in pharmaceutical dosage forms that is accurate, easy, and fast. The separation is achieved using a mobile phase of 50:30:20 v/v acetonitrile, methanol, and potassium dihydrogen orthophosphate buffer, with pH 4.0 adjusted with orthophosphoric acid (10%) on a Phenomenex-luna C18 (250 mm * 4.6 mm, 5) column. We have used a flow rate of e 1.2 mL min−1, with UV detection at 283 nm. Nebivolol and Cilnidipine have retention times of 2.37 and 7.69 min, respectively. For both Nebivolol and Cilnidipine, a linear response is seen over the concentration ranges of 2–10 g/mL (R2 = 0.998) and 4–20 g/mL (R2 = 0.997). For Nebivolol, the limit of quantitation (LOQ) and limit of detection (LOD) are 0.13 and 0.40 g/mL, respectively, and for Cilnidipine, they are 0.11 and 0.35 g/mL. The percent recovery for Nebivolol is 101.4–101.7%, while for Cilnidipine it is 100.5–100.9%. It is found that the procedure is accurate, exact, linear, sensitive, and robust.

The COVID-19 pandemic has increased everyone’s exposure to the Internet thus there has been an addition of new people who now access the Internet and its applications. Hence, the quantity of big data has grown tremendously in the era of smart city life. Thus concern for security of big data has risen. The attacks currently faced by big databases such as misuse, misrepresentation, modification, and unauthorized users and such attacks have increased in number. To ensure the security of big data stores to prevent misuse, misrepresentation, modification, and unauthorized users issues related to insecurity in storage units and not rendering it vulnerable to attackers, it is needed to enhance the secrecy, privacy, and increase the capacity for hiding secret cover. While cryptography guarantees authentication, integrity, non-repudiation, etc. In this paper, the author has studied popular data hiding techniques, especially steganography and cryptography used for provides security to big databases.

Generative Adversarial Network is an emerging technology and research area in machine learning from the time 2014. Generative Adversarial Networks (GANs) is a novel class of deep generative models that has recently gained significant attention. GANs learn complex and high-dimensional distributions implicitly over images, audio, and data. Though carry countless exhilarating prospects, these solicitations likewise increase the mindfulness of the hazard of bogus imageries which may basis enormous destruction. This work highlights some key GAN forms, problems, and research gaps in this study. In addition, we address the benefits that GAN could provide to humans as well as potential solutions. To end with, centred on the perceptions enlarged, we extant encouraging study advices in this hastily emergent arena.

The increasing rate of personal automotive usage within the urban areas is a result of the aggressively growing economy, dilapidated policies, and subvention are the primary causes creating concerns for transport and automotive parking. The coordination between parking policies and traffic management is disclosed; however, parking is turning into a barrier to the through-traffic operation. Also, it is chargeable for the inefficient use of accessible resources, even the choices are created on ad-hoc basis while creating policy. The objective of this research paper is to study the parking choice of behavior made by the user and enhance the experience, making it effortless for the user to park the vehicle. This study integrates these aspects and presents the progressive review of models and studies on the parking system. The methods used to determine and research uses are contextual inquiry, data collection, flow models, surveys, usability testing using SUS and RTP. The designed HCI-based user interface, inside the car dashboard showed a substantial enhancement in parking choice of behavior made by the user and increase in environmental sustainability factor. The decreased parking finding time resulted in increased productivity of the user.

Natural language processing is one of the fascinating areas of artificial intelligence to comprehend human language. It has been widely used to address various issues to understand human language by computers. It has been utilized to pre-process unstructured data and investigate authoritative records to help in legitimate decision-making. The majority of data found in the Indian legal system are unstructured. Various courts disperse legal documents by genuinely dissecting the data of particular cases and getting the choice from perspectives decisions and the rule regulation. Legal professionals’ traditional way of analysing, and making decisions is time-consuming. Hence, in this paper, the authors have focused to fabricate a text pre-handling procedure on authoritative reports to eliminate undesirable texts. The authors have proposed a methodology to pre-process legal documents for content characterization, and text arrangement and examine their specialized commitments to find the major words found in legal documents. Natura language processing techniques like bag of words, count vectorization, etc., help to examine the catchphrases of the court procedures after removing the unwanted texts from legal documents to extract major words found in legal documents.

In this pandemic, people have not been able to celebrate their achievements and accomplishments with their friends and family. One of those achievements is the convocation/graduation ceremony of the students who have successfully invested their hard-work and dedication for achieving this prestigious milestone. The number of people that can attend a physical graduation ceremony is limited due to restricted venue space, seating, and food. Students are unable to invite all of their family members and friends to their big day. Students can invite as many family members as they like to a virtual convocation event. Their family may be anywhere in the world, but they would not miss this momentous occasion. No one is left behind, whether it's kids, parents, or their family and friends, thanks to live transmission to remote audiences and on-demand post-event sessions. Students often return to their native country and are unable to attend their graduation ceremonies. This enables them to attend and create memories with their friends, classmates, and family. This is where the project idea comes into play, where a cross-platform app will be made that will be able to provide a real-life environment to the students and their family members where they will be able to enjoy the convocation ceremony without breaking any social distancing norms.

Oil spill is a type of pollution which affects both directly and indirectly the human beings, economy of the nation, and the marine life. Oil on top of the ocean damages numerous aquatic organisms since it stops sunlight that is sufficient in achieving the surface of the ocean and lowers dissolved oxygen levels. Oil-coated birds and marine mammals can perish hypothermia because crude oil destroys the insulation. In addition, ingested oil is poisonous to affected creatures and habitat. The answer towards the above-mentioned issues is towards building a methodology employing deep learning towards precisely recognizing oil spills and oil-like spill from ocean for taking appropriate action. So, we in this paper have deployed different pre-trained deep learning models towards classification of oil spill in ocean. In addition, different deep learning models performance are compared and validated in terms of accuracy and losses for proposing the best deep learning model for classification of oil spill in ocean.

The roof garden irrigation and drainage system is a take on the traditional roof gardening to ease the strenuous effort and help the environmental balance. As we know while more and more buildings are getting constructed, people are losing valuable places for gardening so to overcome this, people are now doing roof gardening which is a hobby that is also good for the environment and to make people more encouraged to do roof gardening, we proposed this system. The task has been accomplished using an Arduino Uno (ATmega328P), a soil moisture sensor communicating the primary moisture status of the soil, two water pumps for irrigation and drainage each, and NodeMCU for wireless control. This device tends to help in the daily irrigation of multiple tops/surfaces at the individual required time and drain excess water which is harmful to the plant and environment if kept unattended. This device also opens a platform for further studies on, for example, the water that is drained can be used to see the number of harmful products that products contain and how this water can be harmful to surroundings and plants.

The highest rate of accidents occurs on the highway roads during the night. Most of the drivers use high beams while driving at night which causes blurred vision of the person coming from the opposite direction. This intense headlight beam hits the driver’s eye causing him to lose perception and ability to see objects or movements that are not directly in line with their eyes. To avoid such accidents due to temporary blindness of the drivers, we can use a smart system to reduce high-beam glare. The system will detect the high-beam glare using the sensor and will automatically reduce the glare so that the driver does not get distracted and maintains a clear eye vision while driving at night. This system is important because the accident rate is much higher at night only because of the high-beam glare of vehicles from the opposite direction.

One of the greatest challenges in today’s world is to satisfy the energy needs of a growing population, in a sustainable way. In Developing countries, actual electrification of rural areas is hindered by high costs and unreliability of service, which often discourage rural communities to avail energy services even when they are available. This study aimed at better understanding the energy struggle of Indian rural communities, and propose efficient technological solutions to address it. Taking the case of a community located in the Himalaya region, a field based, participatory and bottom-up approach is taken to assess the community needs, expectations, and actual resources. A prototype of solar panel using machine learning-assisted automatic sun tracker system is proposed to provide sustainable and affordable electricity to the community.

IoT is a network of things or objects that can share information using the Internet. IoT is a set of things connected via the Internet to communicate with each other. As part of wireless sensor network, a set of sensor nodes are known as motes. To accomplish a specific task, different types of motes are existing. In the era of IoT, a wide variety of motes are available for different IoT applications. In the research area of IoT, vast number of motes are available with variety of capabilities. But, due to the scarcity of overall practical knowledge, it is difficult to decide which type of mote should be considered appropriate for a particular application and how to evaluate performance parameters of IoT devices. This paper discusses in details about different types of motes and gives comparative study about the same, for IoT application development. This study will enable one, to identify suitable mote for their application and its performance evaluation. Cooja emulator is used for performance evaluation of different types of motes. It supports a large number of different mote types for IoT applications. The comparative study of specifications is based on motes available in the Cooja emulator.

Livestreaming platforms are discernibly the most comprehensive sources of data in real time. Such websites enable users to broadcast content like the games which they are playing, while providing them the opportunity to interact with viewers watching the livestream. Twitch.tv is one of the most popular livestreaming platforms across the globe with millions of monthly active streamers and viewers. Owing to the COVID-19 pandemic, there has been a shift in the conventional lifestyle of the people, with them turning towards online alternatives like Twitch.tv for leisure. This change has led to an increase in the engagement of users in these livestreaming platforms by manifolds. Concurrently, a lot of data is generated from this sudden inflow, which can prove very useful in understanding the general consensus of the crowd. This data is very important, and there is a need to construe the true emotion of the people in real time, which is reflected in the comments made by them in the chat section of livestream. The streamers on Twitch.tv can consequently refine their content immediately based on the feedback that they can infer from the responses given by the users. But, due to the sheer volume of data and convoluted nature of the chat due to the use of emojis, emotes, and emoticons, there are bound to be inconsistencies, human errors, and other esoteric references which are exceedingly complex to dissect, making the task of language processing difficult and leading to incoherent results. Taking into account the hindrance posed by these issues, we have taken up the task to achieve fairly accurate emotion prediction by putting forward machine learning and deep learning techniques. This will involve the creation of a labelled dataset that can be used for training and evaluating the algorithms. Given how context-specific most comments are on the platform, this will be an extensive task. The project will also require the creation of an end-to-end system that performs emotion analysis and giving results in real time through feedback-loops.

Due to the increase in many tools, the relevance of data analytics in health data systems has been growing tremendously as a result of this exponential increase. As a result, it has also led to renewed confidence in the development of data-oriented frameworks for sensor learning in health information systems. With funding provided by artificial neural networks, researchers are looking into the possibility of creating a flexible machine that redefines the recent vision of artificial intelligence. As computing power and rapid processing of reality have increased, the rapid return on technology has often been compromised in terms of conceptual control and production of highly functional and semi-relevant input details that are mechanically designed tails. In this paper, we provide an in-depth analysis of the relative relevance and potential downsides of the approach and its attitude as well as an advanced examination of fundamental learning in health information systems. Additionally, the paper discusses the challenges, security, and protection in healthcare systems with the use of machine learning approaches.

Present paper proposes an antenna-based metamaterial waveguide for broadband applications which works on 7.8 GHz for its return loss. Analysis has been developed X band waveguide which covers coupled split ring resonator having broadside version and C-band waveguide to excite the metamaterial structure as a waveguide to coaxial transition. This structure is numerically analyzed using commercial software CST studio tool. Structure is simulated on dielectric substrate RT duroid with dielectric constant 2.33. Obtained return loss and VSWR are −35 dB and 1.2, respectively.

We all tend to use the native language as much as possible. Even though English is the most popular and professional language, merely 20% of the population can communicate using it. In this planned project, We have developed a secure messaging application with a translation feature considering privacy issues of the user. The goal of this application is to provide messaging functionality with a translation feature (translate the message from English to Hindi and vice versa), maintain privacy by blocking screenshot/s and screen recording, encrypting and decrypting message/s. To implement this, I have used Android OS as it is used by nearly 70% of the world’s population. We have also used the firebase platform and Google ML kit. Google ML kit helps to achieve translation features.

Analysis of textual data helps to understand the perception of people by studying the various senses of words in the text. The sense of a polysemous word varies as per the context in the sentence. The technique for determining the correct interpretation of a polysemous word according to context is called as Word Sense Disambiguation (WSD). Recently, researchers have proposed many algorithms to solve this linguistic ambiguity problem in different languages. In this paper, we give a general summary of current trends in WSD in terms of automation of disambiguation approaches. We also mention the challenges and future directions for research for WSD systems. We also propose a system based on these future directions for research, which may increase the accuracy of WSD system for Indian languages.

One day in the month of November, 2019, the world received a major setback when it understood that a new pandemic called COVID-19, or the Novel Coronavirus had taken over to create havoc among the people. It was first started in a wet market of a small province in China. After that it has spread all over the world like a bonfire. Many countries got under its grip, namely USA, South Korea, and Italy where the situation was totally out of control. During the last two years, India suffered huge loss due to COVID-19 in terms of life, property, and other assets. India is the second largest most populous country with 130 crores was highly affected due to COVID-19. Out of all the aspects, education was the worst affected sector. There was no option left but to implement e-learning as a methodology of teaching and learning. It has emerged as one of the major sources of business, like e-commerce, learning methodology, and e-learning. E-learning is a methodology of teaching and learning where the teacher teaches using multimedia, and the learner learns using the digital mode of education. This mode of teaching and learning has indeed brought a revolution in the education process because neither the teacher nor the student needs to be together in one place. There are numerous subjects which can be taught online, ranging from technical to non-technical subjects. Literature is an imitation of fiction or non-fiction. Online could be the best mode of instruction for literature students. Therefore, this paper is an attempt to make an analysis of the implications of e-learning in education, and its implementation to teach literature by the teachers.

Identification of inter-ictal activity has always presented as a diagnostic challenge, for neurologist consuming much of their time. The automation of the process can provide the required support to the neurologist. Publically available Bonn data dataset has been used for this work. We have created two second segments of public data and created its scalogram which acts as an input to our model, whereas earlier researchers have worked on complete 23.6 s data. LeNet-5-based model is used as classifier. The goal of this work is to distinguish inter-ictal activity with and without presence of various artifacts. Accuracy of 98.03% has been accomplished for the public dataset.

Stock expense assessing is a notable and a huge point in money-related and academic assessments. Protection trade is an unsteady spot for expecting since there are no tremendous standards to survey or anticipate the expense of the stock in the protections trade. Various strategies like particular examination, fundamental assessment, time series examination and quantifiable assessment, etc., are used to predict the expense in tie monetary trade; yet, these strategies are by and large not exhibited as a dependably OK assumption instrument. In this paper, we done a long-short term memory (LSTM) method for managing and expect stock market costs. LSTMs are really completed in deciding stock expenses, returns, and stocks showing. We outline the arrangement of the LSTM with its noteworthy features and versatile limits. We revolve around a certain social occasion of limits with a fairly basic impact on the stock expense of an association. By taking assistance of assessment, it was discovered that the limit marks of the new article and that supported gauging definite result. Though monetary trade can never be expected with hundred percent precision due to its equivocal space, this paper targets exhibiting the capability of LSTM at assessing the stock expenses.

A microstrip antenna array with a resonant frequency of 28 GHz for satellite applications is built, and its radiation characteristics are investigated in this research. The design is a microstrip patch antenna array fed by a corporate network. The Feko Suite was used to examine the radiation properties of the samples. The design was optimised with the help of the optimisation feature in Feko Suite in order to get better radiation properties. The need for satellite applications to operate in the 28 GHz frequency spectrum is also emphasised. An in-depth discussion of the design of a microstrip antenna array as well as the explanation of the results is provided.

River network extraction is crucial to keep track of the water resources. Various methods have been implemented in times series to yield profound and incisive outputs and are still being developed and combined with predefined available methods. We have carried out a structured survey on these methods and have presented them with their outputs. There are numerous Web sites available for data set collection. Some generalized methods are available like image processing, using predefined models, or developing user-defined algorithms. For image processing, various segmentation methods are available out of which clustering- and threshold-based segmentations are mostly used. Predefined models such as CNN, ResUNet, YOLO, Faster CNN, and MSCFF are available. These algorithms can be used for the extraction of river networks but might not yield higher accuracies. Hence, this paper concentrates mainly on approaches for the extraction of river networks from satellite images.

Managing attendance is a vital task for every institution. Considering the COVID pandemic where many organizations have resorted to online mode of working, it has become imperative to maintain social distancing and digitize various processes. Thus, for maintaining attendance of the students of schools/colleges or employees of a company, a touchless attendance system is required that records the attendance by capturing faces and does not waste time. This one-of-a-kind application uses a client–server model and captures the faces of students/employees through video feeds from mobile phone cameras, and the images are sent to a server, where image processing is used to process the faces. Further, with the help of dlib and the face recognition library, it identifies the faces and records the attendance in the software itself. The processed image is again sent back to the client android application, and the user gets notified about their attendance. Additional functionalities for data analysis and updating data have also been added to the system. Thus, the whole attendance system is an effort to make the attendance activity easy and efficient.

Defect discrimination is a crucial factor in evaluating the quality of mango. The flaw mark on the flesh remark about the defunctness of mango fruit. The internal defect of mango fruit is often directly proportionate to the area of flaw marks on the skin. The defect discrimination of mango fruit manually with respect to the area of flaw marks is challenging, erroneous, and time-consuming. Hence, the development of automated techniques for defect discrimination of mango fruit is needed. Here, an image processing method is developed to discriminate the defective flaw mark on the skin of mango fruit into three categories, i.e., good, average, and bad. The experimentation reveals that the proposed method successfully categorizes the mango as per the area of flaw marks on the mango fruit’s skin.

Rising oil prices, rapid depletion of oil reserves, and environmental concerns are driving demand and penetration of electric vehicles (EVs) in the market. One of the very important part in Electric vehicles is the Battery Management System (BMS), which ensures the EVs reliability and also reduce the battery’s maintenance cost, increase the safety and protection of battery and increase the battery’s life. This paper discusses the various functions of BMS, various challenges and opportunities in the development of BMS to meet the safety and future needs.

The new coronavirus, which produces a highly contagious sickness, enters the picture. Globally, the coronavirus disease (COVID-19) has boosted the use of Internet commerce. It has resulted in an increase in the number of FTUs, or first-time e-commerce users, in India, who were previously unable to purchase online. Customers are depending on Amazon more than ever before in their social isolation and self-quarantine attempts, according to Amazon, one of the country’s and the world’s major e-commerce businesses ( https://retail.economictimes.indiatimes.com/re-tales/impact-of-covid-19-on-onlineshopping-in-india/4115 ). The goal of this study is to determine the elements that influence customers’ online purchases of mobile phones, as well as how that practice has altered since the global COVID-19 outbreak. An online survey was done, and data were gathered from primary sources to interpret the objectives. The goal of the study is to figure out how successful online advertising is at raising awareness and what the link is between Internet advertising and purchasing decisions.

The cloud-based healthcare system has opened up new possibilities in the medical profession. By monitoring the whole digestive tract, endoscopic technology has improved the diagnosis of gastrointestinal illnesses and diseases. Image compression enhances the frame rate, which helps the diagnostic procedure. The aims of presenting this paper are to put light on brief review of image compression into two categories intra- and inter-prediction for endoscopy video compression, which is efficient and simple. Our research work shows an acceptable compression performance; the initial phase of work is an implementation of intra-prediction, and the second phase of work is inter-prediction technique, which shows a mark able performance in term of PSNR value, bit rate and compression ratio.

Virtual reality (VR) is known for providing immersed user experience of 3D virtual world. In this paper, we have evaluated the learning of the engineering students and have also compared their learning with other the students who have not used VR. To conduct our study, we have developed a 3D model for various engineering fields such as, architecture, automobile and mechanical and so on. Further, we have also included VR interactions such as grabbing various 3D model, changing materials of the model and so on, to provide students an immersed experience of various 3D models of their respective discipline. Through the study, we have observed the students approximately 90% of students had good learning experience and have understood the concept better.

With the everyday advancements in information communications technology (ICT), many technological developments have colossally affected education, and one such example is augmented reality (AR). The goal of this research paper is the visualization of the difficult concepts by utilizing the AR technology, which can be a means to innovative learning for the kids seeking elementary education. The paper highlights the importance of utilizing the idea of incorporating learning through AR, instead of letting the kids watch videos on screen. In the current paper, the innovative learning results have been shown picturizing a simple concept of the solar system working. Here, free 3D models of the planets are downloaded from Sketchfab, thereby converted into.glb form, imported, implemented, and developed in PlugXR software. Such applications of AR can envision creative minds to an ever-increasing extent of learning, and unlike the YouTube videos that provide the students with the knowledge, these AR experiences allow for the kids to remain active in their real surroundings.

Security and trust of any electronic system is one of the major concern in the present era of globalization. The malicious or unexpected modifications capable enough of accessing the system to change its functionality and weaken the system: Hardware Trojans. However, a methodology that is reference-free, proficient, and experience high-false positives is greatly encouraged. The study involves: a reference-free hardware detection analysis, gradual-N-justification methodology, and local outlier factor. GNJ methodology is an extensible linear algorithm that produces a list of suspicious signals, detects the HT, and reduces the false-positive rates. Local outlier factor adds a depth to the study by classifying the data points based on the local density, reduces the suspicious signals, and subsequently the number of iterations required to bring out the most suspicious signals. The proposed methodology does not fail to bring out any non-maskable Trojan if inserted into the circuit. Therefore, the proposed methodology extracts the most suspicious signals from a list of suspicious signals with high accuracy and less time.

In this paper, an IoT-based wireless sensor network for automatic irrigation of agricultural fields has been designed and developed. Four numbers of sensor nodes are placed over the desired locations to collect data from an agricultural field, and a controller node is placed near the pump to control the water flow of the agricultural field. The sensor node consists of a microcontroller NodeMCU ESP8266 and a soil moisture sensor, and the controller node is made of NodeMCU ESP8266 and a relay to control the on–off state of the pump. The soil moisture sensor of these nodes collects soil moisture content data and delivers it to the cloud or server using the message queuing telemetry transport (MQTT) protocol, through wireless communications modules. The ThingSpeak cloud platform is used where the data is aggregated and processed to make a decision about the status of the pump. This proposed system ensures that the agricultural farm receives the exact amount of water it needs and prevents water waste. As the moisture content in the soil drops below the specified value, the system will automatically start the pump. Detected parameters and current pump status are displayed on the user’s Android application. The farmer uses the Android app to receive updates on the status of his field.

Handwritten document contains a lot of unreadable texts or elements which have no meaning such as struck out words or characters. If such words are fed into a handwriting recognition system, there can be a drop in the accuracy of the system, and it may result into predicting false words. In this paper, we propose an approach to detect these struck out words. With the help of CNN model, we train the model to recognize and differentiate the normal words from the struck out words. For this purpose, some common types of struck out strokes were handled. In order to obtain a handwritten text free of these words, this method may identify strike-through text, locate the word/character, and erase these words. The model was trained on a set of English words and characters that we generated, and it was then put to the test on a range of texts that contained words that had been struck out. The experimental results demonstrate the accuracy of the proposed approach, where the models achieved accuracy levels of 100%.

The demand for food is increasing every day because of the growing population. However, the proportional growth in the production of crops is not happening. This problem can be addressed using information technology in the agricultural process. In this paper, an Internet of Things-based (IoT) system has been designed and developed to monitor soil health. The chemical and physical properties of the soil can be evaluated by measuring different parameters contained in the soil such as nitrogen (N), phosphorus (P), potassium (K), temperature, moisture, and salinity as well as electrical conductivity. Three sensor nodes are developed using NodeMCU ESP8266, NPK sensor, EC sensor, soil moisture, and temperature sensor. First, the data obtained from the procured sensors have been tested and validated in the laboratory by adding different amounts of urea and water. These sensor nodes, which are powered by solar, are placed at some desired location in the agricultural fields to collect the soil data. All the sensor data are then sent to the cloud over a Wi-Fi network using Message Queuing Telemetry Transport (MQTT) protocol for monitoring, further analysis, and decision-making. The data of the soil parameters are stored in a cloud platform, which will allow monitoring of the data in graphical and numerical formats. An android application is developed for real-time monitoring of the data.

Solid-state transformer (SST) is essential equipment in electrical devices. Because of its compact size, it is often employed in high-power applications. Converter-based solid-state transformer has various features such as low weight and size, high frequency, and no core losses. These capabilities make a solid-state transformer replace the power distribution transformer. Bidirectional power flow is also possible in the SST, so it can be easily integrated with the utility grid and renewable energy storage or larger batteries. SST can be made up of three stages: rectifier, dual-active bridge (DAB) converter and inverter. DAB converter is a vital component of SST and is used for power transfer from one circuit to another circuit. Two h-bridges with eight IGBT switches and one high-frequency transformer (HFT) serve as isolation in the proposed DC-DC DAB converter. As an application of SST, a DAB converter was studied and simulated in a MATLAB Simulink environment and phase-shifted pulse generated from dSPACE 1103 platform. This work employs the single-phase shift (SPS) control approach to provide steady output and power transmission.

In the present age, job satisfaction of employees is considered very important because it is a major contributor to an organization’s productivity and profitability. Many leading and reputed surveys indicate that HR managers find it difficult to identify the various factors of employees’ job satisfaction. This paper aims to provide the major factors of employee job satisfaction and also a predictive model that predicts the changes in level of employee work productivity when changes are made to the values of factor(s) of employee job satisfaction. In the first phase (factor identification), 50 factors were obtained from our survey and were subject to factor analysis for dimension reduction. As a result, 50 factors were reduced to 14 major factors or 22 independent variables which were clubbed to 14 major factors. In the second phase, a predictive model is formulated by using the multiple regression technique. The model has 78.8–84.7% accuracy with 95% confidence level. This model helps the HR managers to predict the work productivity of employees within the organization if the value of the factors affecting job satisfaction is increased/decreased. For this part of the project, the newly extracted 22 variables which were clubbed to 14 factors during the factor analysis were taken. In summarization, this study will help HR managers to identify and reduce the employee dissatisfaction thereby aiding with the employee retention and the work productivity of the employees within an organization.

‘Mahabharata’ is the most popular among many Indian pieces of literature referred to in many domains for completely different purposes. This text itself is having various dimension and aspects which is useful for the human being in their personal life and professional life. This Indian ‘epic is originally written in the Sanskrit Language. Now in the era of natural language processing, artificial intelligence, machine learning, and human-computer interaction, this text can be processed according to the domain requirement. It is interesting to process this text and get useful insights from Mahabharata. The limitation of the humans while analyzing Mahabharata is that they always have a sentiment aspect toward the story narrated by the author. Apart from that, the human cannot memorize statistical or computational details, like which two words are frequently coming in one sentence? What is the average length of the sentences across the whole literature? Which word is the most popular word across the text, what are the lemmas of the words used across the sentences? Thus, in this paper, we propose an NLP pipeline to get some statistical and computational insights along with the most relevant word searching method from the largest epic ‘Mahabharata’. We stacked the different text processing approaches to articulate the best results which can be further used in the various domains where Mahabharata needs to be referred.

Every year thousands of people die around the world in motorway accidents, and one of the main reasons for this is drivers’ drowsiness and fatigue. According to a survey by the Central Road Research Institute (CRRI) in 2019, drivers who exhaust themselves doze off, while driving are accountable for about 40% of road mishaps. To reduce the road mishaps, a system to monitor driver’s alertness by detecting the visual features of the driver by finding the drowsiness state of the driver is proposed. It deals with an algorithm which considers the frequency of the eye-blink called PERCLOS, that make use of the eye coordinates obtained from Dlib’s Haar cascade model to determine eye’s state of the driver either open or close and sounds an alarm if the driver is found to be in drowsy state, the warning can be deactivated manually rather than automatically. This algorithm performs better than current drowsiness detection systems in both accuracy as well as speed at adequate lighting conditions. The frames captured of driver are of 640*480 resolution at over 20 fps to determine drowsiness of the driver and give accuracy of 98%. It is also affordable as it does not require any expensive hardware, only a built-in Android camera is required to provide a warning sound when the proposed system predicts that the driver is drowsy. This research result can serve as an important component in ADAS, and it can ensure safety of drivers and minimise financial and personal losses caused by accidents.

Cooling towers are mainly utilized to disperse the heat of thermal power plants (TPP). The availability of cooling tower is directly proportional to the maximum availability of TPP. To ensure the maximum availability of cooling towers, a mathematical model is developed followed by optimization using four swarm intelligence-based metaheuristic algorithms, viz. Grey Wolf Optimizer, Grasshopper Optimization Algorithm, Dragonfly Algorithm, and Whale Optimization Algorithm. The Markovian birth–death process and Chapman-Kolmogorov differential–difference equations are utilized to derive the objective function of availability associated with the proposed model. It is observed from the numerical investigation that the Whale Optimization Algorithm performs better than all other metaheuristic algorithms in providing the optimized values of various failure and repair rates and predicting the overall availability of the cooling tower.

Imbalanced data distribution implies an uneven distribution of class labels in data which can lead to classification bias in machine learning models. The present paper proposes an autoencoder-based sampling approach to balance the data. Concretely, the proposed method utilizes a conditional variational autoencoder (VAE) to learn the latent variables underpinning the distribution of minority labels. Then, the trained encoder is employed to produce new minority samples to equalize the sample distribution. The results of numerical experiments reveal the potency of the suggested technique on several datasets.

In present situations, most people are suffering from insomnia. Due to stress, nowadays sleeplessness has a negative impact on human health. The medication treatment for this illness causes many side effects. Further, this leads to anxiety, depression, irritation, etc. It affects the performance of the user who is currently experiencing insomnia resulting in adverse conditions; it may even lead to accidents. This proposed work mainly discusses a non-pharmacological approach that reduces insomnia and provides better sleep which reduces side effects and implements better sleep. The latest diagnosing factors depicting the illness, even though it is lagging with certain conditioning factors, will be able to overcome those deviations from the existing technology. Therefore, this provides a betterment in the healthcare technology with the latest advancements for both the user and end-user applications. The results have provided the best results as it can be further developed as a wearable device in the future for the healthcare industry.

This paper describes the design, development, and implementation of a standalone bioinformatic tool for the prediction of putative prophage loci in bacterial host genomes using statistical measures, based on the algorithm published as the “Prophage Loci Predictor for Bacterial Genomes” and described as the loci predictor algorithm. This algorithm proposed a novel approach to the problem of detecting prophage regions in bacterial genomic information using particle swarm optimization, using a fixed size pattern lookup table to detect virus-like pattern distributions in the host/bacterial genome. As this algorithm was designed with the intension of providing highly consistent and fast performance, the time-to-process sequence is the primary metric for evaluating the performance of the tool, and the processing speed was expected to scale only with the size of the genome under consideration and not on the size of the pattern database as is the case with other algorithms in its class. The implemented tool was evaluated using both the algorithms test and training sets and was shown to obtain a linear co-related performance as expected in both training and prediction phases of the performance testing.

The Internet of Things (IoT)-based third eye glove is the solution for the blind and employs ultrasonic sensors and an Arduino UNO board. The goal of the IoT is to link physical items, such as computers, to other physical objects, such as other people. Currently, it is a market-enabling technology that is quickly evolving and increasing. As many as 1.6 million youngsters are amongst India’s 40 million blind people. Blind people have a tough time travelling on their own. They have to rely on others in many aspects of their existence. When they’re strolling along the street, this is the biggest problem. With a stick in hand, they will not be able to see every obstruction. One may wear this sophisticated glove design for a long time. The third eye glove will assist the blind individual in achieving their goal. Sensors, microcontrollers, and buzzers are all integrated into the glove’s design, which is IoT. The sneaker emits a buzzing sound as the user approaches an obstruction. Smart gloves, which have sensors and can cover a larger area, are being developed to increase productivity. Smart health monitoring and smart gloves ensure that the user does not have any problems.

Designing peripheral warnings and notifications to support teaching-learning is progressively increasing. However, existing work usually fails to effectively integrate real-time alerts and tackle poor students in a blended classroom. We present an in-class multimodal alert method for teachers and students to address the challenges. The system utilizes performance prediction and classification of students for real-time alert. The classification of students based on course performance helped in optimizing the number of alerts. The peripheral device selection aided in preventing the disruption in the lecture follow. Moreover, alert content delivery timing (start, during, and end of the class session) is used to reduce alert fatigue. We reported the design and the initial study results. The results show that 25 teachers and students reacted positively to the system design, technology, and features.

Image classification is commonly utilized in computer vision tasks like remote sensing, scene analysis, object detection, and image retrieval. We use the Sentinel-2 satellite dataset in our study to classify land cover using the convolutional neural network (CNN) method and determine different plant indices, water indices, and geology features in the Pichavaram mangrove forest. Mangrove woods protect the shoreline from ocean waves, Tsunami storms, and soil erosion. They are efficient at sequestering and storing carbon and mitigating climate change. It is critical to map the extent of the mangroves to protect them. We need an automation solution because the geographic expansions are so large. For automatic feature extraction, CNN is used. Our datasets have been divided into water bodies, vegetation, and marshy plains. Sentinel-2 has a variety of uses, including land monitoring, yield prediction, land cover, flood volcanic eruption detection, and landslide detection.

The ultimate aim of preoperative medical assessment is to minimize the patient’s surgical and anesthetic preoperative morbidity or death, and to return him to normal functioning as soon as feasible. Any preoperative evaluation must begin with a thorough history and physical examination, with an emphasis on risk factors for cardiac and pulmonary problems as well as determining the patient’s functional ability. We analyzed a real-world dataset of patients’ blood tests by using machine learning techniques and data science. We had raw data in CSV format; we had to normalize the data and then analyzed the data and then visualized the data. After that we used different machine learning models on the data. We have used different libraries such as sklearn, pandas, scipy, numpy, and matplotlib. Among a large variety of algorithms, we have used by the included study; the highest accuracy we have obtained is 83.34% from logistic regression technique using random state. We have obtained close to similar result using decision tree with 81.81% accuracy. As far as our other algorithms used, we have obtained 73% accuracy and 80% by using gradient boosting classifier and support vector machine algorithms, respectively. The remaining algorithms that were not fruitful enough giving us −13.3% and −9% for gradient boosting regressor and linear regression, respectively.

With the steep growth in information technology and its global reach, as well as the common citizen’s ever-increasing reliance on technology, data privacy and security have become a major source of concern for individuals all over the world. In today’s era, computing devices like virtual servers, databases, physical servers, databases, and many more devices are occupied with confidential data. This paper is an exploratory case study that analyzes the various algorithms and methods proposed across the various domains, and a comparative analysis was done.

Spatial data processing had been a dominant contributor in a wide assortment of applications including health care, urban planning and infrastructure designing. As the complexity of data has increased due to higher dimensions, diverse approaches have been successfully tested for nearest neighbor queries, clustering algorithms, etc. These algorithms have been optimized and adapted for CPU, GPU and hybrid models as combination of CPU and GPU. Closest pair (CP) computation is a frequently used operation in range and distance-based queries. Although there are different algorithms and optimizations available for CP computations, suitability of GPU for this operation is yet to be explored. Hereby in this paper we propose CPU-GPU Hybrid Model, to optimize the closest pair (CP) of points problem concerning two-dimensional floating point values. As part of the combined model, the sorting phase is implemented on CPU and computation of closest pair is implemented on GPU. In this work, three CP algorithms are analyzed on CPU and CPU-GPU combinations. With the proposed algorithm on the CPU-GPU combined model, up to 12 times speedup is achieved in the closest pair computation time. Several parameters were tuned to work with the device-specific features and the overhead in terms of data transfer is also analyzed.

Different cervical pap smear cell categorization schemes have recently been presented, the majority of which were binary classifications of normal and abnormal cells. This paper presents the findings of a comprehensive investigation on machine learning and deep learning algorithms for binary and multi-class classification on pap smear images from the Herlev dataset. There are 917 photos in this collection, divided into seven normal and pathological categories. The Google Colab platform was used to generate models utilizing all of the techniques using scikit learn and the keras library from TensorFlow. To begin, several repetitions of processes such as feature importance selection, data normalization, standardization, PCA, T-SNE, and others have been imposed on models such as SVM and XGBoost in this work for machine learning approaches. Second, it was demonstrated in this work that a transfer learning-based CNN model from deep learning can outperform machine learning models in terms of binary and multi-class classifications. Furthermore, it was discovered in this work how computationally time efficient it is to apply a transfer learning model, which required roughly 25 min for 100 epochs. Finally, with several iterations of processes and outcomes, this work demonstrates that given enough data for a multi-class pap smear image classification system, the transfer learning CNN model has a higher potential to get the best results than the machine learning models used.