VOLUME 10, ISSUE 8, AUGUST 2021

Abstract: Sentiment Analysis (SA) is the process of determining the opinion of a text written in a natural language to be positive, negative, or neutral towards any specific target such as individuals, events, topics, products, organizations, services, etc. SA has its challenges, and one of them is sarcasm. Sarcasm is a form of communication that is intended to mock or harass someone by using words with the opposite of their literal meaning. It is often used to express a negative message using positive words. However, Sarcasm detection is one of the most challenging tasks in Natural Language Processing (NLP) especially for the Arabic language which has a rich nature and very complex morphology. It has gained relevance recently, due to its importance in improving the performance of various NLP applications including SA. In this paper, we propose an approach for automatic sarcasm detection in the Arabic text of Twitter data by using the Support Vector Machine (SVM) classifier to classify sarcastic tweets based on different N-gram features and using several weighting schemes. The experimental results obtained are promising. The best results by SVM classifier for all feature sets and several weighting schemes achieved overall accuracy equal to 86.60%, which these results are quite high especially regarding Arabic text.

Keywords: Automatic sarcasm detection; Sarcasm; Sentiment analysis; Text Mining; Support Vector Machine (SVM); Arabic language

Abstract: Business Analytics is the emerging domain of the 21st century. Machine learning algorithms control a growing range of business functions once governed by humans, including business intelligence. Most BI products go further than just enabling data aggregation and reporting. They may also provide insights or optimization suggestions using predictive analytics functions. In this paper we start with data acquisition. Any acquired/ given data can be analysed and conclusions drawn accordingly. The acquired or given data usually exists in its crude or raw state. Data pre-processing helps to format the data into useful form by removing redundancy and noise, eliminating missing and non-numerical values, and also by normalization. Data analysis and visualization are carried out to improve the statistical analysis of given data. Logistic regression is carried out on the data since it contains lot of columns with categorical values. Accuracy, precision, and f1 score of the model have been measured. Various conclusions can be drawn from this interdependent data set and can be stored as historical data for future analysis. Linear Regression is also carried out on the data set and r-squared values noted. R-squared is a statistical measure of how close the data are to the fitted regression line. A ML model is built by employing both logistic regression and linear regression for the automobile industry. This Business Intelligence model is a boon to the manufacturers and sales department in identifying their product in the 21st century market

Keywords: Business Analytics (BA)/ BI (Business Intelligence), Machine Learning, Data pre-processing, Logistic regression, accuracy, precision, and f1 score, linear regression, data analysis and visualization, R-squared, Business Intelligence.

Abstract: Quality of Service (QoS) refers to ability of a network to provide improved service to selected network traffic over various underlying technologies including Frame Relay, ATM, Ethernet and 802.1 networks, SONETS and IP-routed networks.The traffic through this network may include data types such as email, file sharing or web traffic. Other forms of traffic include voice and video. These are considered as sensitive forms of traffic. They often require guaranteed or regulated service. In this paper, we deal in detail with two factors affecting QoS: Congestion and Queuing. Identifying and rectifying the above problems to reduce packet loss, latency and jitter on the network is the errand for the day. Few of the Queuing Algorithms like First in First out (FIFO), Priority Queuing (PQ), Round Robin and Weighted Round Robin (WRR) are explained in brief pictorially. The internet is often vulnerable to attacks from possible hackers who try to compromise the system in order to illegally poach the resources of the system under question. These attacks are famously called web attacks and are a very common problem amongst the computer fraternity. Though there are several existing systems to counter the problem of attacks on the web, most of these systems have their own drawbacks, as in they do not provide classification on any other grounds except frequency, thus causing many web attacking http requests to fall out of the bracket. The objective of our project is to detect these web attacks from the http requests based on many parameters, and classify them as web attacks or not.

Keywords: Frame Relay, ATM, Ethernet and 802.1 networks, SONETS and IP-routed networks, Congestion and Queuing, First in First out (FIFO), Priority Queuing (PQ), web attacks, web attacking http requests.

Abstract: Object Detection and tracking of objects is a field that has many applications in this rapidly developing society with more cameras being set up all over the world. Traffic surveillance has become the most pressing issue in increasingly developing cities. Due to poor traffic management in the city of Bangalore, a lot of manpower and hours are being used up. Our project provides a system that detects and monitors vehicles, pedestrians, traffic signals, and signboards and keeps a count of the number of objects per class passing through. This is built using a custom YOLOv4 dataset and functions pertaining to Bengaluru Traffic and implemented using YOLOv4 and Tensorflow. Our model managed to raise the number of objects detected by over 60% to 94% but went down in predicting accuracy from 90% to 65% compared to foreign datasets.

Keywords: YOLOv4, Tensorflow, Non-max Suppression, Darknet.

Abstract: In this paper, a Machine Learning-enabled intelligent navigation system is presented. It will recommend routes in a road network by minimizing source to destination distance by choosing right shortest path between source to destination , it also take care and avoids categorically marked COVID-19 hotspots. The Q-Learning based system takes the source and destination as inputs from the users and recommends a safe and shorter path for traveling. It reduces the risk of getting exposed to the contaminated zones and contracting the virus by bypassing the red covid19 hotspot zones.

Keywords: Reinforcement Learning, IoT, Intelligent Navigation System, Route Planning, Q-Learning, Covid19 Hotspot

Abstract: Motorcycles are one of the most popular means for transportation. As the popularity and usage of two-wheelers increase, the number of accidents also inevitably increases. Road accidents are one of the primary causes for non-natural deaths. In order to solve this problem, numerous countries have proposed vehicle laws, making helmets compulsory for both the rider and the passengers. Also, the number of people riding on motorcycles is limited to 2. In India, any person above the age of 4 must compulsorily wear a helmet. Even though wearing a helmet is essential and compulsory, not everybody follows this rule, as there are multiple instances of people not wearing a helmet while driving a two-wheeler. There are also several instances where people do triple riding or riding with more than the allowed number of passengers. To mandate this, we have created a model using OpenCV, TensorFlow and YOLO to identify rule violations. The model takes the front view image and side view image of the vehicle and using object detection techniques, it identifies riders with and without helmet. The model also checks for multiple rider rule violations. If any rule is violated, the licence plates of such riders are automatically extracted and stored

Keywords: Automatic Number Plate Recognition (ANPR), You only look once (YOLO), Helmet Detection, Person detection, Machine Learning (ML), Optical Character Recognition (OCR), Common Objects in Context (COCO)

Abstract: The Jaya algorithm is one of the recently developed innovative, algorithm-specific parameter-less optimization algorithm. This study presents a short review of exploration and exploitation approaches used by researchers to improve the performance of the basic Jaya algorithm. The objective of this paper is to present collectively the important strategies adopted to modify the basic Jaya algorithm focusing on exploration and exploitation only. This study considers the recent journal publications about the Jaya algorithm and its improvement. It is observed that researchers have focused on the solution update equation of the basic Jaya algorithm to improve the performance and balancing the exploration and exploitation in the modified Jaya algorithm.

Keywords: Exploration, Exploitation, Jaya algorithm, Strategies to balance exploration and exploitation.

Abstract: Vehicle insurance processing using images is a critical sector with a lot of room for automation. In this study, we look at the topic of car damage detection. Vehicle damage detection. Using images taken at the site of an accident can save time and money when filing insurance claims, as well as provide more convenience for drivers. Artificial Intelligence (AI) in the sense of machine learning and deep learning algorithms can assist in solving problems. A vehicle-damage-detection technique based on transfer learning and a mask regional convolutional neural network (Mask RCNN) are utilized to quickly handle accident compensation problems. The algorithms identify the damaged section of a car, determine its position, and then estimate the severity of the damage. Very satisfactory results have been produced using transfer learning to take advantage of available models that have been trained on a more generic object identification challenge.

Keywords: Car Damage Detection, Machine learning, Prediction, Mask RCNN, Transfer Learning, Deep Learning

Abstract: In order to cope with the changing education system and the evolving new technologies, it is important for a student to identify his field of interest and select his best among the available wide range of courses. Many students opt the courses which are not of their interests as they will not be having much knowledge about the courses of their interests. Most of the students in our society decide their future based on what their elders say or they rely on their friends or their family and does the same course which they had done or doing. There will be no proper guidance for them on choosing their subjects or courses. This project is a part in progress of education towards better course recommendation. We use a machine learning program that asks the client questions, and recommends the better stream based on the skills and academic performance provided. The program also serves as a data collection platform to support the drive for more data on course recommendation.

Keywords: Career recommendation, Machine Learning, Recommendation system, Hybrid approach, Form based.

Abstract: Data Analytics of COVID-19 is not just enough for curbing down this deadly disease which now has become a pandemic. The detection of Severe Acute Respiratory Syndrome corona virus 2 (SARS cov-2), which is responsible for corona virus disease 2019 (COVID-19), using chest X-ray images has life- saving importance for both patients and doctors. [1] X-rays are cost-effective and widely available at public health facilities and hospital emergency rooms, they could be used for rapid detection of possible COVID-19-induced lung infections. Therefore, towards automating the COVID-19 detection, in this paper, we propose a viable and efficient Deep Learning Based Chest Radiograph Classification (DL-CRC) framework to distinguish the COVID-19 cases with high accuracy from other abnormal and normal cases. A unique dataset is prepared from four publicly available sources containing the chest view of X-ray data for COVID-19, pneumonia, and normal cases. Our system consists of Convolution Neural Network (CNN) architecture capable of detecting masked and unmasked faces also which will be dealt with in the next paper. The encouragingly high classification accuracy of our proposal implies that it can efficiently automate COVID-19 detection from radiograph images to provide a fast and reliable evidence of COVID-19 infection in the lung that can complement existing COVID-19 diagnostics modalities.

Keywords: Convolution Neural Network (CNN) architecture, COVID-19, Severe Acute Respiratory Syndrome corona virus 2 (SARS cov-2), deep learning based chest radiograph classification (DL-CRC) and radiography images.

Abstract: The objective of this project is to generate power continuously from the renewable energy resources. Here the renewable energy resources used in this project are solar power and rain water. Now-a-days using solar energy for the generation of electricity has become very popular. Particularly in the agricultural fields the usage of solar panels for electricity increasing day by day. The flaw in this concept is whenever there is sun, there will be solar energy production. But if there is no sun light i.e., during rainfall sufficient power will not be generated. For overcoming this problem this project is needed. This project is based on the principle of using hybrid mechanism of combining solar power technology with piezo electric power technology In this project, we will have hybrid panel i.e., one side it will have solar panel and other side it will have piezo electric plate. Based on the weather conditions, the plate will be rotated automatically using dc motors. So whenever the day is sunny, solar panel will be faced up and whenever there is rainfall, the piezo electric plate will be faced up. For sensing the sun, we are using an LDR and for detecting the rainfall we are using moisture sensor.

Keywords: hybrid energy harvester, photovoltaics, piezoelectricity, electromagnetism.

Abstract: One of the theoretical challenges going through scholars is to increase an accepted definition of white-collar crime. The predominant attribute is that it is financial crime dedicated through a individual of respectability and high social fame in the route of an occupation. While Edwin Sutherland’s theory of white-collar crime has enlightened ,penologistic, and administration researchers, the thought may additionally human ecology stressed attorneys ,judges and lawmakers. One motive for this confusion is that white-collar crime in Sutherland’s lookup is each a crime committed by a specific kind of person, and it is a unique type of crime. Later lookup has indicated, as applied in this book, that white-collar crime is no particular kind of crime, it is solely a crime committed by means of a specific type of person.

Keywords: Convenience idea • Criminology • Definition • Edwin• Social repute

Abstract: Diminishing of vegetation is the main issue in Indonesia as a country with second largest forest in the world. It invites the increasing of global temperature and climate change. In local scale (province and regency level), the conversion from vegetation into another land type could lower the quality of life in such area. To prevent the decrease of vegetation, local governments have been trying to stop the conversion but it is difficult to ensure in daily activity with million people in those areas. Therefore, the objective mean is needed to monitor the numbers of vegetation in a region, e.g. satellite imagery and geographic information system. Satellite imageries can be used to analyse land cover change, especially the conversion from vegetation to other land covers. In this study, a land change modeler using two different dates of map was implemented after hard classification using iterative self-organising clustering in IDRISI Selva software. Decreasing vegetation trend and the direction of change can be seen after comparing two dates of land cover classification map.

Keywords: Satellite Imagery, Multispectral, ISOCLUST, Hard Classification

Abstract: In medical imaging, segmentation plays a vital role towards the interpretation of Xrays, CT Scans and MRIs where salient features are detected and extracted with the help of image segmentation. Finding an optimal medical image reconstruction methodology is becoming increasingly difficult as technology advances. As a result, medical imaging has benefited from advancements in analysis and diagnosis. Without undergoing surgery, clinicians and radiologists employ various modalities ranging from X-Rays and CT-Scans to ultrasonography, and other imaging techniques to visualise and examine interior human body organ and structures. The focus of this study is on the segmentation approached applied to chest x-ray images, tumour obtained from CT and MRI images. Keyword: Pattern Recognition, Image Segmentation on X-rays, Tumour Detection, MRI, Medical imaging.

Abstract: Different path loss models have been predicted for different locations. Nevertheless, none of these models can be regarded as a superior model, because environmental factors play a vital role in the path loss of every environment. In this paper, signal attenuation prediction models for Global System for Mobile Communication (GSM) networks in the University of Nigeria, Nsukka for four different networks namely Airtel, Globacom, Mobile Telecommunication Network (MTN), and 9mobile networks were proposed. Field measurements based on the signal strength and path loss of GSM operating at 1800MHz were carried out for the development of the proposed attenuation model, in the area for the four GSM networks. The measured data for signal strength and path loss were used to develop the models. To formulate the proposed attenuation models for the considered networks in the area, the data collected during field measurements were analyzed using linear regression analysis. The proposed models were compared with the measured and four popular standard attenuation models such as Hata, Cost 231-Hata, SUI, and ECC-33. The path loss for the standard empirical models was gotten from simulation using a standard MATLAB 2016b package. The results showed that the proposed attenuation models performed better than all the considered models based on its least error value.

Keywords: GSM, path loss, attenuation models, linear regression, received signal strength.

Abstract: Due to the successful emergence of the internet of things, sensor-based smart door using Temperature Sensor and Oximeter. A usable non-contact IR temperature sensor that can measure the body temperature without any physical contact is implemented. This paper describes a working prototype of Smart Door using Temperature Sensor and Oximeter system using MLX90614 temperature sensor, DIY Arduino with Oximeter MAX30100 where Ultrasonic Sensor is used for the distance measuring. In this prototype sensor, data is acquired and analysed to give proper feedback to the person with its temperature and oxygen level. The sensor vitals are collected and sent to the Arduino using shielded cable i.e., through wired communication, respectively. Analysis of a person's vitals based on ambient Temperature and Pulse gives a person's real-time temperature and Pulse condition so that if the condition is not normal, then the buzzer and red LED will flash so that preventive measures can be taken to avoid further complication. Per user, data can be saved in the system database for further reference.

Keywords: DIY Arduino, LED, Oximeter, prototype, Smart Door, Temperature Sensor.

Abstract: Mobile computing are one of the most trending culture of today’s Information age to utilize the computational power of micro and macro devices such as mobile or other hand held devices with full fledge of technological rich computational tasks with ease to perform in a very accurate and reliable manner. But to connect these all micro devices with proper communicative coupling orientations in the form of mobile networking structure are really becomes a very typical and challenging tasks, because these all micro devices contains various utilities to connect them from one to another such as Bluetooth Technology, WI-FI Technology, but to use these utilities some limited number of devices have to be connected with each other and a very less amount of data have to be transfer at a minimum time instance. To achieve the productivity and maximum through put the need of the hour is that we have some potential mobile networking mechanisms which works as similar as computer networks and utilize all resources with full of its computational capacity and becomes a computationally rich. In this paper we proposed a ad-hoc networking model among various micro and macro devices such as mobile or other hand held devices in a small area spectrum generally called a Campus Area Network (CAN) using the full potential of mobile computing architecture and these type of networking possible among the mobile or some other hand held devices. We also proposed a new protocol name says MCC (Mobile Communication Configuration) with its header along the sizes of each parameters or modules which presently existed in the MCC header and also represent the packet structure with its data capacity size. Because this networking model works as a Local Area Network (LAN). These type of networking model suitable for small size of area such as offices, universities, institutions, campus etc. The detailed mechanisms of each module of this networking model are presented in the next upcoming sections of this paper.

Keywords: Mobile Networking Technology, Mobile Computing based Networking, CAN network of mobile computing, Mobile communication using MCC protocol.

Abstract: Every automatic water distribution system (AWDS) fails partially or completely at some point or points during its lifetime. Measurement of reliability for a AWDS under various failure conditions is necessary. Numerous techniques have been developed for estimating AWDS reliability. It was found that types of failure (mechanical, hydraulic, and water quality failure) and techniques of reliability analysis are mixed together inappropriately. Water is the basic need of all living organism and human mankind, without water living is impossible. In recent days the rapid population growth causes insufficiency and wastage of drinking water which leads to scarcity of water and uneven distribution of drinking water. Next issue is that the supplied water is sucked more by individual home unit using suction pump which leads shortage of water to the remaining houses in the locality. In this paper a system has been mode led to overcome the above stated problems. The main aim of this paper is to distribute only required amount of water needed, thus ensuring there is no wastage and block in supply of water. In order to implement the proposed system each home unit must be provided with water flow sensor and water flow switch which is controlled by arduino mega board. Flow sensor generates series of electric pulse through which water utilize by the user, flow rate and the amount of water supplied can be calculated.

Keywords: Arduino Mega, Water flow sensor, Valve, Water flow switch, LCD.

Abstract: Fake news in the media is not new. It has been with us since the development of the earliest writing systems. Fake news have caused a lot of damage to humanity and hence the need to detect it. The term “fake news” is not new but detecting it quickly has really been a problem. This study used random forest and decision tree algorithms on a dataset containing both fake and real news to do classification. The software used for the experiment was Weka and the result generated show that random forest correctly classified instance is 100% and incorrectly classified instance is 0% while the decision tree correctly classified instance is 93.6364% and incorrectly classified instance is 6.3636%. The results is a proof that random forest algorithm is a better classification tool as compared to decision tree.

Keywords: Fake news, Random Forest, Decision Tree, Algorithm, tool, Classification.

Abstract: Information Technology plays a significant role in the current era. The ease with which business operations are done has made Information Technology to be embraced widely. It brings with it low cost of operations, and a low learning curve. The education sector has benefited from Information Technology (IT) significantly, especially use of mobile phones in learning. The main objective of this paper is to represent a systematic literature review on M-Learning in Kenyan universities with the underlying aim of investigating the challenges facing mobile learning in Kenyan Universities. The findings point to the need to focus on improving the ease-of use and effectiveness of mobile learning in Kenyan universities. Systematic literature review is one of the most commonly used method to review technology content in information technology and computer science since 2006. In total, 250 articles have been collected that were written between 2005 and which were published in the EBSCO, Elsevier, Google Scholar, Science Direct, and, ProQuest One Academic were screened. After the articles were screened and selected, 50 were selected to be included in the study. The data were assessed to identify the independent and dependent variables in the study that help the researcher to identify the challenges of using mobile phones for learning in Kenya Universities. Key words: University, m-learning, challenges, student, mobile technology

Abstract: Human face recognition is a field of study in Computer Vision, and is also considered a research area of Biometrics (similar to fingerprint recognition, or iris recognition). In terms of general principles, facial recognition has a great resemblance to fingerprint recognition and iris recognition, but the difference lies in the specific extraction step of each field. While fingerprint and iris recognition has reached maturity, which is widely applicable in practice, facial recognition remains challenging and remains an interesting area of research with many. people. Compared to fingerprint and iris recognition, facial recognition has a richer data source (you can see human faces in any photo or video clip related to people online) and requires less more controlled interaction (to perform fingerprint or iris recognition, human input requires cooperation in a controlled environment). Currently, the face recognition methods are divided into many directions according to different criteria: still image based FR (2D) recognition is the most popular, but the future will probably be. 3D (because the layout of many 2D cameras will give 3D data and give better and more reliable results), it can also be divided into two directions: doing with image data and doing with video data.

Keywords: Human faces, iris recognition, facial recognition, fingerprint recognition

Abstract: To build flexible systems that work in a variety of lighting conditions and run on mobile phones or handheld PCs, robust and efficient face detection algorithms are required. Appearance-based methods are mainly employed to achieve high detection accuracy. They solve a two-class problem by using a probabilistic framework or finding a discriminant function from a large set of training examples. To solve this problem, it is necessary to find more distinctive features, which can capture the structural similarities within the face class. In this paper, I’m propose a new feature, called joint Haar-like feature, for detecting faces in images. This is based on co-occurrence of multiple Haarlike features. Feature co-occurrence, which captures the characteristics of human faces, makes it possible to construct a more powerful classifier. The joint Haar-like feature can be calculated very fast independently of image resolution and has robustness against addition of noise and change in illumination.

Keywords: algorithms detection, probabilistic framework, finding a discriminant function, Haar-like feature.

Abstract: Neural Networks (NN) is a subset of Machine Learning and is being used widely these days in predictive analysis of pre-processed images. CNN (Convolutional Neural Network) is a popular NN algorithm and it clearly outperforms ANN in this project. Inception V3, ResNet50, MobileNet and Xception [1] are the existing CNN models but are found to be less accurate and more time consuming. In our R&D lab we have developed a new CNN model called the H5 model. It is the best fit after the output is obtained from Haar Cascade Classifiers. A model which was developed for facial detection and distinction is now used for all objects detection with more accuracy focusing on five regions with different pixel Intensity scheme. The encouragingly high classification accuracy of our proposal implies that it can efficiently automate COVID-19 detection from radiograph images to provide a fast and reliable evidence of COVID-19 infection in the lung that can complement existing COVID-19 diagnostics modalities.

Keywords: H5 Convolutional Neural Network model, Convolution Neural Network (CNN) architecture, COVID-19, Severe Acute Respiratory Syndrome corona virus 2 (SARS cov-2), deep learning based chest radiograph classification (DL-CRC), Tensorflow, Haar Cascade Classifiers, different pixel Intensity scheme, facial detection and distinction.

Abstract: Business Process Redesign (BPR) is the complete overhaul of a key business process with the objective of achieving a quantum jump in performance measures such as return on investment, cost reduction and quality of service. BPR is basically resolves the old form of organization through the improvement in order of magnitude. Professionals in the business and academics have developed a number of methodologies to support this competitive rearrangement that forms the current point of convergence, many of which have not been successful. This paper suggests the use of Data mining as a method to support the process of redesigning a business by extracting a lot of essential knowledge hidden in bulk extent of data maintained by the organization through the Data Mining models.

Keywords: Data Mining, Knowledge Management, Business Process Redesign, Business reengineering, Artificial Neural Networks.

Abstract: Product codes are usually applied in high data rate wireless communication systems to achieve good performance. The product code composed by the simple block code can be decoded by a decoding method based on the syndrome-norm. Syndrome-norm decoding method uses lookup table to roughly determine the type of the error pattern and then take specific corresponding decoding method. The scale of the lookup table will increase with the rise of the bits of the maximum errors since it requires to store all the possibilities of the error types. However, the existing pattern library formation method is high computational complexity and fail to form a pattern library with error bits above six. This paper proposed a mathematical model for fast generating a library based on the iterative expansion of the error patterns, which makes it possible to shorten the computational complexity in comparison with the known forming approaches and support higher error correction capability.

Keywords: product code decoding, iterative formation, syndrome, norm, syndromic-norm decoding, library of error patterns.

Abstract: One of the challenges in today’s world is to deal with fake/false information on the internet or simply on social media. social media was created with intention of connecting people over the internet but after time passes it becomes the mainstream data and information transfer system because of its accessibility and availability. Social media is indeed the mainstream to broadcast news, information, Social Status on a lightning-fast basis but it takes a great cost to deliver information with less credibility. create chaos over the mainstream of information which is social media apps ex-Facebook, Instagram. in this paper analytics of social media platforms like Twitter, Facebook, Instagram, WhatsApp included. as it’s free to create profiles on social media the problem of fake profiles and their impact on people was also analysed.

Keywords: Social Media Analysis, Fake Profile/News, Network,data integrity and security, challenges of social media, fakeness detection, and information integrity

Abstract: Credit risk is the probability of a loss resulting from a creditor’s failure to repay a loan or fulfil any other contractual obligations towards the investor. Traditionally, it relates to the hazard that a lender may not receive the owed head and premium, which follows a disruption of incomes and expanded expenses for collection. Unnecessary cash may be written to create additional income to cover for credit risk. Despite it is being impossible to know exactly who will default on commitments, satisfactorily surveying and overseeing credit risk can diminish the seriousness of a loss. The lender or investor earn a bonus for risking credit default and lending money in the form of interest from the borrower or issuer of a debt obligation. When lenders or banks provide mortgages, credit cards, visas or various types of credit or loans, there is a hazard that the borrower is probably not going to reimburse the loan. Likewise, if an organization provides credit to a client, there is a hazard that the client is not going to pay their solicitations. Credit risk additionally clarifies the risk that a guarantor may stall to make payment when asked or that an insurance company will be unable to pay a claim. Credit risks are determined based on the borrower’s general ability to reimburse an advance as indicated by its unique terms. To assess credit risk on a consumer loan, loan specialists inspect the five Cs: credit history, capacity to repay, capital, the loan’s conditions, and associated collateral. Banks have been the most important institutions of money lending and deposits. Primary functions include accepting deposits, offering loans, credit, overdraft, providing liquidity and discounting of bills. Secondary functions include providing safe custody of valuables, loans on valuables, corporate and consumer finances. Though the structure of banks has remained the same, the functionalities have been boosted. Automated tools, bots and computers have modernized the banking system. The dataset accumulated over a period of time is so huge that, automation tools and computer programs are the need of the day. In this paper we have tried to enhance the present bank credit-debit system by the use of Artificial Intelligence. Machine learning is a subset of AI and directly trains the machine by feeding the historic and runtime data collected during transactions. The machine which is trained is now capable of taking decisions, thereby making predictions. This would characterize the dataset as stored and predicted outcomes. Every business enthusiast would have keen interest to carefully study the performance of a financial institute for his/her benefit. In this assignment we have used both classification and regression algorithms to create a ML model of prediction. Linear regression model is designed from scratch using formula method. Classification algorithms like Support Vector Machine (SVM), Random Forest Classifier and KNN algorithms are effectively applied to fit to the dataset. Comparisons must be made during implementation to understand the pattern of predicted data. Regression algorithms like linear regression (developed from scratch) will be a boost to the accuracy of the assignment (categorical data excluded).

Keywords: accepting deposits, offering loans, credit, overdraft, providing liquidity and discounting of bills, Automated tools, bots and computers, Machine learning, Support Vector Machine (SVM), Random Forest Classifier and KNN algorithms, linear regression (developed from scratch) , historic and runtime data collected during transactions, AI, five Cs: credit history, capacity to repay, capital, the loan’s conditions, and associated collateral.

Abstract: Vehicular Ad Hoc Networks (VANETs) are self-organizing, self-healing networks which provide wireless communication among vehicular and roadside devices. Applications in such networks can take advantage of the use of simultaneous congestions, thereby maximizing the throughput and lowering latency. In order to take advantage of all radio interfaces of the vehicle and to provide good quality of service for vehicular applications, we developed a seamless flow mobility management architecture based on vehicular network application classes with network-based mobility management. Now-a-days accidents are mostly caused by delay of the driver to hit the brake or by the negligence by the driver.

Keywords: VANET, simultaneous congestion, mobility management, brake.

Abstract: Progressive Web Apps (PWAs) and WebGL have emerged as influential technologies in modern front-end engineering, revolutionizing the way web applications are developed and experienced by users. This paper explores the significance of PWAs and WebGL in the context of front-end development, highlighting their key characteristics, benefits, and real-world applications. By examining the synergies between PWAs and WebGL, we discuss the opportunities and challenges associated with their integration. Through a comprehensive literature review, we analyze the current state of research and identify future trends and opportunities in this field. The paper aims to provide valuable insights for front-end engineers, researchers, and stakeholders interested in leveraging PWAs and WebGL to create immersive, performant, and engaging web experiences.

Keywords: Progressive Web Apps (PWAs), WebGL, Front-End Engineering, Web Development