This research is designed to develop an advanced framework for the prediction of diabetes and the personalised recommendation of drugs. This research proposes a multi-attention enhanced task incremental learning coupled gradient boost deep neural network (MA-TIL-GBNN) approach aimed at the prediction of diabetes and recommendation of drugs based on their types. The approach integrates MA techniques into the DNN and generative adversarial network-based data augmentation (GDA) to effectively analyse the health indicators. Additionally, the TIL enhances the training of data, and the Light GBM facilitates efficient processing. The experimental results using the Diabetes Health Indicators Dataset showed 97.79% accuracy, 98.22% sensitivity, and 97.36% specificity, respectively.

The study introduces a machine learning-based drug recommendation system that is targeted to help achieve better care for diabetic patients. Using PCA, eight key components were identified to capture the most significant variance, and six unique patient clusters were formed by unsupervised learning. A Random Forest model-based recommendation of medication for individual patient profiles was developed at the user level, providing an accuracy of 0.86. While expert clinical input would always be invaluable, this system demonstrates how data can support and refine treatment strategies, giving healthcare professionals an evidence-based tool for managing diabetes.

Tuberculosis (TB) is a serious infectious disease that remains a global health challenge. While chest X-rays (CXRs) are widely used for TB detection, manual interpretation can be subjective and time-consuming. Automated classification of CXRs into TB and non-TB cases can significantly support healthcare professionals in timely and accurate diagnosis. This paper introduces a hybrid deep learning approach for classifying CXR images. The solution is based on the CoAtNet framework, which combines the strengths of Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs). The model is pre-trained on the large-scale ImageNet dataset to ensure robust generalization across diverse images. The evaluation is conducted on the IN-CXR tuberculosis dataset from ICMR-NIRT, which contains a comprehensive collection of CXR images of both normal and abnormal categories. The hybrid model achieves a binary classification accuracy of 86.39% and an ROC-AUC score of 93.79%, outperforming tested baseline models that rely exclusively on either CNNs or ViTs when trained on this dataset. Furthermore, the integration of Local Interpretable Model-agnostic Explanations (LIME) enhances the interpretability of the model's predictions. This combination of reliable performance and transparent, interpretable results strengthens the model's role in AI-driven medical imaging research. Code will be made available upon request.

The software development life cycle (SDLC) is incomplete without the software testing phase. It's the act of checking that a piece of software really does what it's supposed to. Test case creation is one of the testing tasks that have a major impact on the quality and speed with which the process is completed. Research into the automated production of test cases has been extensive because of the time and energy it can save over the human method of creating test cases. While most of the recommended methods are based on UML models, other publications have given a specifications-based method of creating test cases. This literature analysis focuses on automated test case generation strategies based on use case specifications and the techniques used to verify them. The analysis also highlights the ways in which the methods diverge when used to solving certain pressing problems in software testing.

Computational Storage Devices, also known as In-Storage computing or In-Suit Processing, offer higher computing power than traditional storage devices. Innovation in computational storage devices is more important than ever because of the exponential growth of digital data produced by both individuals and organizations. With the advent of non-volatile storage devices in mainstream computing (e.g., hard disk drives and solid-state drives), the idea has gained wide appeal, with several academic and commercial prototypes becoming available. This survey aims to provide a systematic overview of the work being done in the area of computational storage and to indicate future directions. This overview considers and examines a number of research questions to comprehensively summarize, analyse and discuss various storage devices such as Hard Disk Drives (HDD), Solid-State Drives (SSD) and Computational Storage Devices (CSDs) in terms of design, programming model, acceleration and energy efficiency. Also, the Load Balancing algorithm has been reviewed from different domains such as cloud computing, data centers and the Internet of Things (IoT). This review not only focuses on the existing literature problems but also requires some effort to explore the future in this area.

This research introduces an innovative job recommendation platform designed to optimize the job search process through advanced technology integration. The platform features a sophisticated Resume Analyzer leveraging NLP and ML to meticulously extract candidate information. Coupled with a Recommendation System that employs mixture models such as DRNN-BES and collaborative filtering, it delivers personalized job suggestions tailored to individual profiles. Additionally, the system utilizes state-of-the-art web scraping technologies to continuously update its database with diverse and current job listings from various online sources, ensuring a wide range of opportunities are accessible. This integration of technologies not only refines job-matching precision but also significantly improves the efficiency of connecting job seekers with appropriate opportunities. Initial experimental results underscore the enhanced effectiveness and accuracy of our recommendation system, particularly when incorporating user behavioral data into the collaborative filtering algorithm. This paper demonstrates the transformative potential of integrating AI into job seeking, offering a scalable, unbiased, and time-efficient approach to modern employment challenges.

When a borrower doesn't keep their end of the bargain, there is a danger of financial loss known as credit risk. Even though there are numerous factors that contribute to credit risk, careful consideration when approving a loan may lessen the likelihood of the development of non-performing assets (NPA) and fraud. While there are organizations that offer credit rating and credit scoring services to banks for a fee, the researchers continue to investigate different machine learning strategies to increase the accuracy of credit risk assessment. In this study, a novel approach based on a modified version of SVM (Support Vector Machines) is proposed. The proposed methodology incorporates a custom-trained variant of SVM. A function calculates class weights based on the inverse frequency of each class in the training set. The hinge loss is calculated using a function that takes the feature matrix, target labels, weight matrix, and regularization parameter as input. The hinge loss function penalizes the model for misclassifying instances. Further, this work compares three machine learning (ML) algorithms-SVM, XGBoost, and Random Forest with the proposed approach to comprehend the efficacy in spotting credit risk. The proposed custom SVM model is trained with new weights that were found using the current dataset, which provided an accuracy of 96.92% that is higher than the pre-existing models. Also, the findings showed an fl-score of 0.90, a precision value of 0.915, and a recall value of 0.885.

Software Reliability (SR) is a substantial aspect of software quality and one of the most important variables to consider when evaluating the quality of a software product. Software managers have a difficult task in selecting the optimum SR model for a real software development project. Throughout the software life cycle, it is concerned with the creation and maintenance of dependable software systems. Software measurement, software modeling, and software enhancement are 3 steps of a three-step method for boosting software dependability. Each of these stages is critical to establishing a solid software system. To control software quality, it is critical to improving accuracy in predicting dependability. A variety of measures for assessing software dependability have been suggested in the literature. Machine learning (ML) methods have been demonstrated to be effective in assessing many aspects of software dependability. To capture the many aspects of a software system, numerous machine learning approaches have been developed in this work. In this review paper the results of the various algorithms are compared. ML classifiers are compared with each other based on accuracy, precision, and recall parameters to determine the best classifier. Machine learning methods have been used in comparative analysis to estimate the defect level of the software instances. Comparison was placed among SVM, NB, DT, RF, and ANN classifiers. Where ANN shows the best results among other classifiers it has highest accuracy of 65.5%.

Cyber physical systems (CPS) is a ubiquitous concept where objects are connected to the internet and equipped with an ability to sense physical medium and transmit data through network. In recent years, smart home applications have gained popularity due to their various benefits like time-saving, low cost, optimised electricity use. The main component of the CPS is a low-capacity sensor node. Due to resource limitations like low storage space and limited processing power, the well-known security mechanisms used in desktop devices fail to run on these systems. For secure smart home systems, this research paper proposes a lightweight key agreement and authentication mechanism. It protects against various cyberattacks. It also provides communication security, privacy protection for end-users, and a low-cost system for sustainable production. The system's proof-of-concept prototype has been implemented, demonstrating the system's feasibility. The real-world experiments show the system's effectiveness and efficiency in various scenarios.

Heart disease prediction is a challenging task that is under research from many decades. There are several factors that cause heart attacks in patients. These factors can be used to analyse and predict if a patient is having a risk of getting heart attack. This paper presents a risk factor analysis of factors that result in heart attack and put forth the association between different factors. The analysis of the association can help doctors personalize the treatment based on the patient condition. The rules of association, namely support, confidence and lift, have been used to find out how different factors, single and combined, can play a role in causing heart attack to the patients. The risk factor parameters under study are: thal, age, exang, restecg, chol, sex, cp, fbs, trestbps, and thalach. Finding the association between parameters can help us analyse what factors, when combined, can have the highest risk in causing heart attack.

In recent years extensive research is going on for the development of applications which convert physical devices into smart devices. Industry 4.0 adopt the technologies under Cyber Physical Systems (CPS) for the development of such types of smart devices. Increase in the use of such type of smart devices without any security mechanism causes an open invitation for cyber attackers to perform cyber-attacks on such devices. Even current security algorithms are not efficiently work due to some constraints of smart devices. The goal of this research paper is to provide effective solution against different cyber-attacks on CPS applications. This paper proposed session key-based security mechanism which is used for the prevention of cyber-attacks and authentication of cyber devices.

Cyber Physical Systems (CPS) plays an important role in industry 4.0 applications such as smart factory, smart energy, smart transportation, and smart cities. The Supervisory Control And Data Acquisition (SCADA) systems are typically used in Industry environment for CPS. They are industrial control systems (ICS) used in many critical infrastructures such as smart-power grid, refining, as well as infrastructure e.g. water management, oil & gas pipelines, and wind farms.There are five key properties of the Cyber-Physical System (CPS) namely: Design thinking, safety, security, sustainability and resiliency[12] [13].These properties share a common goal – Monitor behaviour of the CPS system and protect them from failure. In order to achieve this goal, it is critical to align the processes and activities of the CPS system to utilize the perspective of design thinking, safety, security sustainability and resiliency. In this research work, we propose a novel framework for CPS system that is motivated from International standards such as ISA (International Society of Automation) 99, ISO (the International Organization for Standardization) 27001, ISO 27002 and Physical and environmental security. The proposed framework aids CPS design, security and safety at early development phases by aligning processes based on ISA-99, ISO 27001, and ISO27002 standards. The framework, Fault-Attack-Safety- Security-Sustainability-Resiliency Tree (FASSSTeR) includes fault trees, attack trees, safety countermeasures, security countermeasures, sustainability and Resiliency measures. The proposed framework can be used during design, development and operation phases for periodic assessment, monitoring and verification of safety and security of CPS system.

The Internet of Things (IoT) has been evolving in tandem with centralized cloud computing to fog computing, edge computing, Semantic computing, etc. Pervasive IoT applications like Healthcare applications generate a huge amount of sensor data and imaging data that needs to be handled rightly for further processing. In traditional IoT ecosystem, Cloud computing ensures solution for efficient management of huge data with its ability to access shared resources and provide common infrastructure in a ubiquitous manner. Majority of the IoT applications are highly sensitive to time and necessities latency bounded execution. There is time delay introduced when the data transmission occurs between the cloud and the application which is unacceptable. This chapter will disclose the various facets and trends in the integration/coupling of evolving computing platforms and disruptive technologies like Fog/Edge Computing, Big Data, blockchain with IoT that overcomes challenges existing in the traditional deployment of IoT environment. Further, there are several issues with IoT and Cloud Computing framework like each individual part of the IoT architecture could act as a point of failure that can interrupt the whole network and secondly, the centralized cloud model is vulnerable to loss of integrity of data. This chapter will review and reveal the trends in bracing computing approaches, managing the huge data using datacenters in Internet of Things ecosystem. The review is advocated with a case study on waste management system applied with the fog/edge computing and cloud computing. Further, various trends in applicability of block chain in the IoT ecosystem is also reviewed. This chapter discusses the fundamental facets of various computing paradigms and approaches that may help to address issues of big data through creating IoT ecosystems.

In the case of the agile software development environment, the key challenge is to generate test cases applying user stories. The newly designed “User story Processor (USP) algorithm” can be used to pre-process the user stories using Natural Language Processing (NLP). This Paper also presents the clear flow for execution for future development. This is ongoing research and as the future scope for real-time run, plug-in with other testing software will be very efficient. Such plug-in can be deployed for agile software tools as an added functionality as each agile project can contain different modules of the project which need multiple types of testing.

In the recent years, the number of people passing away due to heart attack has increased a lot. The lifestyle changes of the 20 century have made people more prone to heart attacks. This paper presents a heart disease classification system using deep learning. The individual parameters causing the heart attack are analysed in terms of risk factor. The risk factor analysis has helped to access the role of each parameter at a personal patient level. The risk factor analysis has led to discovery of redundancy present in the datasets and thereby providing input on how accuracy can be increased. The proposed model used Convolution Neural Network (CNN) to classify the heart data. The UCI heart dataset is used to validate the proposed method. A custom dataset is constructed with new realtime parameters which has been recently discovered. The proposed method has achieved better accuracy when compared to the existing counterparts.

The new concept provoked with genetic and natural characteristics like neuron structure in brain [neural network (NN)] is known as computational intelligence (CI). CI is depending upon three main stakes, namely, NN, fuzzy systems, and evolutionary computing (EC). Fuzzy system utilizes natural language or more specifically language used by human being to represent linguistic ambiguity as well as solves the difficulties in approximate reasoning computation. NN can be defined as computational model designed in line with structure and working of human brain which can be trained and perform job accordingly. This chapter focuses on EC and various facets of EC for machine learning (ML) applications. EC provides a way out to problem of optimization by generating, calculating, and transforming different likely solutions like use of evolutionary programming, genetic algorithms, multi-objective optimization, and evolvable hardware (EH), etc. In the recent years of ML and BigData, computer systems/applications are creating, gathering, and consuming a huge amount of data. Since such enormous data are utilized for ML functions along with applications which are data demanding, rate at which data transmission occurs from storage unit to host unit will be the reason for a bottleneck. Evolvable programming and EH jointly could be a key to solve abovementioned problem of optimization. Solution could be provided by using near-data processing (NDP) or in-situ processing. NDP allows execution of compute-centered applications in situ, which means within or very close to the storage unit (memory/ storage). Outcome of this may be that NDP seeks to decrease costly data movements and thus performance improvement. NDP is made up of a grouping of software system plus hardware system derived from the collection of central processing unit (host unit) plus smart storage system. Smart storage system has various elements namely various disks for storage, cache, frontend, and backend. Now cache memory consists of large capacity which is used for storing hot data (frequently used data) temporarily. The cold data (data used less frequently) are moved forward to underlying storage layer. Frontend is intended for connection between host system and storage system. Backend is intended to support communication among cache and multiple storage disks. One part of backend is backend controller that pedals these disks having capability of Redundant Array of Independent Disks (RAID) and having error finding as well as error-correction methods. This chapter talks about different architectures of NDP which is compatible for ML methods and in which way mentioned NDP architectures boost execution of ML methods. The chapter focuses on developments in recent research in NDP designed for various methods of ML.

Data mining, Theprominent mining problem is the extraction of highly useful sets (HUI) or an extraction utility (UI). The topic of HUI (set of elements of high utility) is mainly the introduction to the set of frequent elements. Mining frequent patterns is a widespread problem in mining, which consists of finding frequent patterns in the transaction databases. Solve the problem of the high utility element set (HUI) with some particular data and the state of the art of the algorithms. To store the HUI (set of high utility components), numerous prominent calculations for this issue have been proposed, for example, "Apriori" FP development, and so forth., yet now TKO required calculations (component extraction sets K in one stage) and TKUs (utility itemsets Top-K extraction) K here TKO is a best stage Top K TKUs is being used. This paper tends to the above issues, proposing another methodology for major HUI k where k is the perfect number of HUI to remove. Extraction of high utility component sets is a remarkable term. Be that as it may, we are utilizing it while we purchase on the web, and so forth. It is a piece of the business examination. The fundamental region of use is the examination of the market bushel, where when the client purchases the thing, he can purchase another to amplify the advantage for both the client and the provider. KEYWORD-Mining useful set of elements of high utility, k mining upper model, upper extraction series of k elements, TKO, TKU. *Corresponding author AshwiniKurhade Department of Computer Engineering, BharatiVidyapeeth Deemed to be University College of Engineering, Pune. Email id-kurhadeashwinibd@gmail.com AshwiniKurhade et al., IJSRR 2019, 8(1), 752-758 IJSRR, 8(1) Jan. –March, 2019 Page 753