The Scientific Temper

A comparative analysis of virtual machines and containers using queuing models

A. Anand — 2024-10-16

Virtual machines (VMs) and containers are two prevalent technologies in cloud computing, each offering distinct advantages depending on the use case. VMs emulate entire operating systems, including kernels, while containers share the host OS kernel, making them lightweight and resource-efficient. This paper presents a novel method for comparing the performance of VMs and containers using queuing models. The proposed method not only provides a more accurate and flexible comparison but also significantly reduces the time required to calculate and perform performance metrics compared to traditional empirical benchmarking and simulation-based approaches. Through this comparison, the paper highlights the conditions under which containers outperform VMs, particularly in modern, cloud-native environments.

AI-driven real-time performance optimization and comparison of virtual machines and containers in cloud environments

A. Anand — 2024-10-16

The accurate calculation and comparison of performance in cloud environments are critical for optimizing resource utilization, particularly with the increasing use of virtual machines (VMs) and containers. This research proposes an AI-driven resource management framework that surpasses traditional machine learning algorithms by enabling real-time, autonomous performance optimization. While machine learning models provide predictive capabilities, they often require manual tuning and retraining for changing workloads. In contrast, the proposed AI-driven system, utilizing techniques such as reinforcement learning and adaptive optimization, continuously adjusts resource allocation based on real-time performance metrics like response time, throughput, and server utilization. This dynamic, self-improving system can respond to fluctuating workloads and network conditions without the need for constant retraining, offering superior flexibility and faster response times. The framework will be validated through extensive experiments across multi-cloud and edge computing environments, demonstrating its ability to significantly reduce calculation time while improving scalability and efficiency. Additionally, this approach incorporates enhanced security mechanisms, combining the isolation benefits of VMs with the lightweight efficiency of containers, providing a comprehensive, real-time solution for cloud-native applications.

Properties on semi-ring of fuzzy matrices with compatible norm

A. Pappa — 2024-10-16

Objectives: Using applying the concept of compatible norm, this study develops an entirely new type of fuzzy matrix semi-ring.
Methods: The main goal of the current work is to present a novel fuzzy matrix idea on compatible norm SFMc.
Findings: This study introduces fuzzy matrices and explains which ones multiply left and right over addition. Addition and multiplication are the two binary operations (+, ⨀) that occur in the semi-ring (S,+,⨀) of the set S.
Novelty: The idea of generalized semi-ring of fuzzy matrix are studied. Using the fuzzy algebra & vector space over [0,1]. Forms a distributive law and comparable of semi-ring fuzzy matrices.

Optimization-based clustering feature extraction approach for human emotion recognition

C. Agilan — 2024-10-16

Human emotions are mental health states that resolve without conscious effort and are followed by physiological effects in the face muscles that represent expressions. In many applications of human-computer interaction, nonverbal communication mechanisms such as emotions, eye movements, and motions are used. Since there is no contrast among the emotions of a face and there is also a lot of variety and complexity, identifying emotions is a difficult process. To model the face, the machine learning system leverages some open features. Automatic emotion recognition based on face expression is a fascinating study area that has been presented and utilized in a variety of fields, including safety, health, and human-machine interactions. Researchers in this subject are willing to develop strategies to understand, code, and extract facial expressions in order to improve computer prediction. Machine learning, being one of the most promising new fields, offers a wide range of applications. In recent years, the support vector clustering technique has gotten a lot of attention. In this research paper, the use of ant colony optimization (ACO) for creating k-cluster planes and assigning each data sample to the correct cluster is proposed in this study as an upgraded clustering approach. SVC is used in this improved technique to refine the clusters created by ACO. The human face expressions are segmented using this upgraded clustering method. The suggested clustering technique is compared to an existing segmentation approach for emotion recognition using a variety of criteria.

Energy-efficient location-based routing protocol for wireless sensor networks using teaching-learning soccer league optimization (TLSLO)

L. Amudavalli — 2024-10-16

Energy efficiency in wireless sensor networks (WSNs) is a crucial and fundamental design consideration. These networks typically consist of numerous small, resource-constrained sensor nodes, frequently placed in isolated or difficult-to-reach areas. This research presents a comprehensive methodology for improving the performance and energy efficiency of WSNs deployed in a designated target area. The research begins with the deployment of sensor nodes equipped with location information and the initialization of critical network parameters. Novel techniques are introduced for efficient node clustering using a Haversine-based K-means Clustering algorithm (HKMC) and an advanced hybrid optimization model, teaching-learning soccer league optimization (TLSLO), for optimal cluster head selection within clusters. Data aggregation at cluster heads is crucial for conserving energy, and data compression techniques, including the novel weighted discrete wavelet transform (WDWT)), are employed to reduce data transmission size. Furthermore, deep learning models in the form of recurrent artificial neural networks (RANN) predict energy consumption patterns, enabling the optimization of node sleep-wake schedules for a prolonged network lifetime. Simulated using Python, the proposed protocol’s performance is evaluated, demonstrating its superiority in terms of energy efficiency, latency, network lifetime, and data delivery ratio compared to existing routing protocols. This research offers a holistic approach to improving WSNs enhancing their efficiency and sustainability in resource-constrained environments.

Dynamic resource allocation with otpimization techniques for qos in cloud computing

V. Baby Deepa — 2024-10-16

Ensuring the quality of service (QoS) in cloud computing environments requires efficient resource allocation mechanisms to manage dynamic workloads and meet user demands. This paper proposes a dynamic resource allocation strategy that integrates gravitational search optimization (GSO) with Harris Hawks optimization (HHO) to optimize resource utilization and maintain QoS in cloud infrastructures. The proposed hybrid approach combines the global search capabilities of GSO, inspired by the law of gravity, with the exploitation and exploration strategies of HHO, mimicking the cooperative hunting behavior of Harris hawks. This synergy enables adaptive and efficient allocation of computational resources based on real-time workload fluctuations, reducing response times, minimizing energy consumption, and preventing Service Level Agreement (SLA) violations. By predicting workload variations and adjusting resource allocation dynamically, the proposed method ensures higher reliability, scalability, and cost-effectiveness compared to traditional resource allocation techniques. Simulation results demonstrate that the GSO-HHO-based approach outperforms conventional optimization algorithms in balancing the trade-offs between performance and resource efficiency, making it a robust solution for maintaining QoS in cloud computing environments.

Evaluating the impact of MOOC participation on skill development in autonomous engineering colleges

R. Chandran — 2024-10-16

The integration of massive open online courses (MOOCs) in higher education has introduced new avenues for skill development and academic achievement. This study investigates the impact of MOOC participation on students’ academic performance within autonomous engineering colleges. Specifically, we examine whether students who engage in MOOCs achieve higher academic outcomes compared to their peers who follow traditional coursework only. A sample of 450 engineering students from autonomous colleges was surveyed regarding their MOOC participation, academic performance, and engagement levels. To analyze the hypothesis that MOOC participation positively influences academic performance, multiple statistical methods were employed. Descriptive statistics provided an overview of student participation and performance trends, while a t-test was used to compare academic performance scores between MOOC participants and non-participants. Regression analysis was applied to determine if MOOC participation is a significant predictor of academic success. Additionally, a Chi-square test examined the association between MOOC engagement and academic achievement. The results indicate that MOOC participation positively correlates with academic performance, supporting the hypothesis that MOOCs can serve as a valuable supplement to traditional education. These findings underscore the potential of MOOCs to enhance learning outcomes in engineering education and suggest that autonomous colleges might benefit from promoting MOOC engagement as part of their curriculum.

Enhancing cloud efficiency: an intelligent virtual machine selection and migration approach for VM consolidation

O. Devipriya — 2024-10-16

Cloud-based computing, despite its numerous benefits, frequently exerts a negative influence on the environment. The primary concern lies in the emission of greenhouse gases and the consumption of electricity by cloud data centers, which demands considerable scrutiny. Virtual machine consolidation (VM) is a widely adopted strategy aimed at achieving energy efficiency and maximizing resource utilization. The consolidation of VMs is a fundamental process in the development of a sophisticated cloud resource management system that prioritizes energy efficiency. The underlying premise is that by shifting VMs onto a reduced number of physical machines, it is possible to achieve optimization objectives, increase the utilization of cloud servers, and concurrently decrease energy consumption in cloud data centers. This proposed solution utilizes the best fit decrease (BFD) approach for VM allocation. An enhanced Greedy selection approach is proposed for VM migration, utilizing the Genetic method optimization method.

Fuzzy logic-driven scheduling for cloud computing operations: a dynamic and adaptive approach

A. Kalaiselvi — 2024-10-16

Cloud computing is a decentralized approach of providing and accessing computer services through the internet. The phrase «cloud computing» is commonly used to describe this setup. The term «cloud computing» refers to a method of running computer programs, data, and services over the internet from a central location rather than on individual users’ local machines. Cloud computing environments face the challenge of efficiently managing and scheduling diverse tasks to ensure optimal resource utilization and system performance. This paper introduces a fuzzy logic-based approach for scheduling cloud computing operations designed to handle the uncertainty and dynamic nature of task execution requirements. The proposed method incorporates fuzzy rules and membership functions to evaluate key parameters such as task priority, resource availability, and execution time. By modeling these uncertainties, the fuzzy logic system dynamically adjusts scheduling decisions to optimize load balancing and minimize delays. The approach offers flexibility in allocating resources and prioritizing tasks in real-time, adapting to fluctuating workloads and system conditions. Experimental simulations demonstrate the effectiveness of the fuzzy logic approach in enhancing system throughput and reducing task completion time, offering a robust solution for scheduling in heterogeneous and complex cloud environments. This method shows promise for improving the scalability and responsiveness of cloud-based operations. Comparisons with three separate scheduling algorithms the first come, first serve (FCFS) algorithm, the round robin (RR) strategy, and the Honeybee foraging (HF) algorithm, show that our method is quite effective. The experimental findings validate the efficacy of our algorithm.

Optimization based energy aware scheduling in wireless sensor networks

M. Prabhu — 2024-10-16

In wireless sensor networks (WSNs), energy efficiency is a critical factor in extending network lifetime, particularly in applications involving multiple target tracking. This paper proposes a novel approach for sleep scheduling in WSNs using ant colony optimization (ACO) to achieve energy-aware scheduling while maintaining high tracking accuracy. The proposed method models the scheduling problem as an optimization task, where ACO is employed to dynamically adjust the sleep and active states of sensor nodes based on their energy levels and target detection requirements. By optimizing node activity, the algorithm minimizes energy consumption while ensuring continuous and reliable tracking of multiple targets. Experimental results demonstrate that the ACO-based scheduling approach significantly enhances network longevity and reduces energy depletion compared to traditional scheduling techniques without compromising tracking performance. This energy-aware solution is well-suited for real-time tracking applications in resource-constrained WSN environments, providing a balance between energy conservation and tracking precision.

Cultural algorithm based principal component analysis (CA-PCA) approach for handling high dimensional data

G. Chitra — 2024-10-16

The exponential growth of high-dimensional data in various domains, such as healthcare, finance, and image processing, presents significant challenges for efficient analysis and predictive modeling. Dimensionality reduction is a key technique to address these challenges, mitigating the curse of dimensionality while preserving the most relevant information. This paper proposes an optimization-based dimensionality reduction approach that integrates principal component analysis (PCA) with cultural algorithm (CA) optimization to enhance the handling of high-dimensional datasets. PCA is employed to transform the data by extracting principal components that capture the maximum variance. However, the selection of an optimal subset of components remains crucial for maintaining model accuracy and computational efficiency. To this end, the cultural algorithm is leveraged to optimize the selection of the most informative principal components by mimicking the evolutionary process of knowledge acquisition in a cultural framework. The proposed approach is validated through experiments on various high-dimensional datasets, demonstrating its superiority in reducing data dimensionality while maintaining high classification accuracy and reducing computational costs. The results highlight the effectiveness of combining PCA with cultural algorithm optimization for dimensionality reduction, paving the way for its application in large-scale real-world problems.

Enhanced otpmization based support vector machine classification approach for the detection of knee arthritis

G. Hemamalini — 2024-10-16

The accurate detection of knee arthritis is essential for effective medical diagnosis and treatment. In this study, we propose an enhanced classification approach using a support vector machine (SVM) coupled with Cuckoo search optimization (CSO) to improve the detection of knee arthritis. The classification challenge lies in tuning the hyperparameters of the SVM, specifically the penalty parameter (C) and the kernel function parameter (γ), which significantly influence the model’s performance. Traditional methods of hyperparameter tuning may be computationally expensive and prone to local minima. To address these challenges, we integrate CSO as an optimization algorithm for the efficient search of optimal hyperparameters. Cuckoo search optimization, inspired by the brood parasitism behavior of cuckoo birds, is applied to optimize the SVM hyperparameters by balancing exploration and exploitation during the search process. CSO efficiently explores the hyperparameter space and finds an optimal or near-optimal solution by minimizing the classification error. The hybrid approach aims to enhance the predictive accuracy and generalization ability of the SVM model. The proposed CSO-SVM framework is validated on a benchmark knee arthritis dataset, and the experimental results demonstrate a significant improvement in classification performance compared to traditional SVM and other optimization algorithms. The proposed model’s ability to optimize hyperparameters with CSO shows promise in achieving higher accuracy, precision, recall, and F1 score, making it a robust approach for knee arthritis detection.

Distribution of virtual machines with SVM-FFDM approach in cloud computing

D. Jayadurga — 2024-10-16

Virtual machine (VM) distribution in cloud computing plays a pivotal role in optimizing resource allocation and improving overall system performance. This study proposes a novel approach for efficient VM distribution using a combination of support vector machine (SVM) and the finest fit decreasing modifier (FFDM) algorithm. SVM is employed to classify and predict resource utilization patterns, ensuring that VMs are allocated based on predicted workloads. The FFDM algorithm, a modified version of the traditional first fit decreasing (FFD) algorithm, is then applied to optimize the packing of VMs onto physical servers by minimizing resource wastage and enhancing load balancing. By integrating machine learning techniques with optimization algorithms, the proposed approach achieves a more effective VM allocation strategy, leading to improved system efficiency, reduced energy consumption, and enhanced scalability in cloud environments. Simulation results demonstrate the superior performance of the SVM-FFDM method compared to traditional VM allocation techniques in terms of resource utilization and operational cost.

An ensemble-based approach for sentiment analysis of covid-19 Twitter data using machine learning and deep learning techniques

M. Jayakandan — 2024-10-16

In the wake of the COVID-19 pandemic, social media platforms like Twitter have become critical channels for public expression, capturing a wide array of sentiments ranging from fear and anxiety to hope and optimism. This paper proposes an ensemble approach for automatic sentiment analysis of COVID-19-related tweets to extract valuable insights from large-scale data. The proposed method integrates multiple machine learning algorithms, including support vector machines (SVM), random forests, and deep learning models such as long short-term memory (LSTM) networks. By leveraging these diverse techniques, the ensemble model aims to improve classification accuracy and robustness in detecting positive, negative, and neutral sentiments. Feature extraction is optimized through natural language processing (NLP) techniques like term frequency-inverse document frequency (TF-IDF) and word embeddings. Experimental results on a publicly available COVID-19 Twitter dataset demonstrate the effectiveness of the proposed approach, showcasing its potential to contribute to public health monitoring, policy making, and understanding of public reactions during crises.

Feature selection in HR analytics: A hybrid optimization approach with PSO and GSO

Jayalakshmi K. — 2024-10-16

In the field of Human Resources (HR) analytics, effective feature selection is critical for improving the accuracy and efficiency of predictive models used for workforce management, talent retention, and performance evaluation. This paper proposes an improved feature selection approach that integrates optimization techniques such as particle swarm optimization (PSO) and gravitational search optimization (GSO) to enhance the performance of HR analytics. By leveraging the exploration-exploitation balance of PSO and the mass-based search capability of GSO, the proposed method efficiently identifies the most relevant features from large and complex HR datasets. The hybrid approach reduces dimensionality, minimizes computational costs, and boosts the accuracy of machine learning models used in HR analytics. Comparative analysis with traditional feature selection methods demonstrates that the proposed technique achieves superior results in terms of prediction accuracy, computational efficiency, and overall model performance. This study highlights the potential of advanced optimization techniques in driving data-driven decision-making processes in HR, offering a robust and scalable solution for managing and analyzing HR data more effectively.

Improving image quality assessment with enhanced denoising autoencoders and optimization methods

V. Karthikeyan — 2024-10-16

In the field of image quality assessment, effective noise reduction is critical for enhancing the perceptual quality of images and improving the accuracy of subsequent analyses. This study proposes an enhancement to denoising autoencoders (DAEs) through optimization techniques aimed at significantly improving image quality assessment outcomes. Traditional DAEs, while effective in reconstructing clean images from noisy inputs, can sometimes fail to adequately preserve intricate image details and structures, which are essential for quality evaluation. Our approach incorporates optimization strategies, including adaptive learning rates, regularization techniques, and advanced loss functions, to refine the DAE architecture and improve its denoising capabilities. By training the enhanced model on diverse datasets containing various noise types and image content, we achieve superior performance in noise reduction. The effectiveness of the optimized denoising autoencoder is rigorously evaluated using standard image quality metrics, including Peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and other perceptual quality measures. Results demonstrate a marked improvement in image quality, leading to more reliable assessments in various applications, including medical imaging, remote sensing, and multimedia content. This work highlights the potential of leveraging optimization techniques to enhance denoising autoencoders, thereby providing a robust solution for improving image quality assessment methodologies.

Some properties of maximal product of two picture fuzzy soft graph

M. Vijaya — 2024-10-16

In this manuscript, the concept of a picture fuzzy soft graph is formally established. Subsequently, we explore the notion of maximal product between two picture fuzzy soft graphs. Additionally, we investigate the complement of the maximal product of picture fuzzy soft graphs, along with the maximal product involving strong and complete picture fuzzy soft graphs. Furthermore, a discussion of various outcomes stemming from this exploration is presented.

Enhanced regression method for weather forecasting

T. Malathi — 2024-10-16

Weather prediction is gaining popularity very rapidly in the current era of artificial intelligence and Technologies. It is essential to predict the temperature of the weather for some time. Traditionally, weather predictions are performed with the help of large complex models of physics, which utilize different atmospheric conditions over a long period of time. These conditions are often unstable because of perturbations of the weather system, causing the models to provide inaccurate forecasts. The models are generally run on hundreds of nodes in a large high-performance computing (HPC) environment, which consumes a large amount of energy. In this paper, LightGBM Regression parameters are tuned by using an optimization technique. Differential evolution (DE) is used to optimize the LightGBM regressor for estimating and forecasting the weather in the fore coming days.

Cloud computing research productivity and collaboration: A scientometric perspective

V. Selvi — 2024-10-16

Cloud computing has emerged as a transformative technology paradigm, revolutionizing the way organizations manage and deliver IT services. As the field continues to evolve, it is essential to understand the trajectory of research and development in the cloud computing environment. This scientometric analysis explores the landscape of cloud computing research, aiming to uncover trends, key contributors, and influential research themes. This work will employ a comprehensive dataset comprising academic publications, patents, and conference proceedings spanning the last five years of data. This research work will identify a significant increase in research output in the field of Cloud Computing over the last decade, indicative of its growing importance in both academia and industry. This work will identify key contributors and institutions that have played pivotal roles in shaping the landscape of cloud computing research. Through the trend topic approach, we categorize research themes within cloud computing, shedding light on emerging areas of interest and shifts in focus. Our analysis also examines the international collaboration network within cloud computing research, illustrating the global nature of this field. This scientometric analysis serves as a valuable resource for researchers, practitioners, and policymakers seeking to navigate the complex and dynamic world of cloud computing.

Improving the resource allocation with enhanced learning in wireless sensor networks

M. Prabhu — 2024-10-16

Efficient resource allocation is crucial for optimizing performance and extending the lifespan of wireless sensor networks (WSNs), which are often constrained by limited energy and bandwidth. This paper proposes an enhanced learning approach (ELA) for dynamic resource allocation in WSNs, leveraging augmented reinforcement learning to adaptively manage energy consumption, optimize routing, and schedule node activity. ELA integrates predictive feedback and real-time data from network states to refine policy decisions, enabling the network to maintain optimal performance under varying traffic loads and environmental conditions. Comparative analyses with existing methods, including deep neural networks (DNN), artificial neural networks (ANN), and support vector machines (SVM), demonstrate that ELA achieves superior results across multiple key metrics: energy consumption, network lifetime, packet delivery ratio, end-to-end delay, and throughput. Our findings indicate that ELA can sustain higher data reliability and throughput while minimizing latency and energy depletion, addressing fundamental challenges in WSNs. The proposed approach presents a scalable and adaptive solution that is well-suited for real-time and large-scale IoT applications, making it a valuable contribution to the advancement of intelligent resource management in WSNs.

Trust aware clustering approach for the detection of malicious nodes in the WSN

Rekha R. — 2024-10-16

Wireless sensor networks (WSNs) are pivotal in a range of applications such as environmental monitoring, healthcare, and defense. However, their decentralized and resource-constrained nature makes them vulnerable to various security threats, particularly from malicious nodes that can disrupt the network’s functionality. To address this issue, this paper proposes a novel trust aware clustering (LEACH) approach integrated with an optimization-based technique for the detection of malicious nodes in WSNs. The proposed model leverages the low-energy adaptive clustering hierarchy (LEACH) protocol for efficient clustering and energy management while incorporating a trust-based mechanism to evaluate the behavior of nodes. Additionally, an optimization algorithm is employed to enhance the accuracy of malicious node detection and improve the overall network performance. The trust model dynamically updates based on node interactions, ensuring that compromised nodes are detected and isolated promptly. Simulation results demonstrate the efficacy of the proposed approach in terms of increased detection accuracy, reduced energy consumption, and prolonged network lifetime, making it a robust solution for securing WSNs against malicious attacks.

Enhanced malicious node identification in WSNs with directed acyclic graphs and RC4-based encryption

Rekha R. — 2024-10-16

In wireless sensor networks (WSNs), ensuring secure data transmission while preventing malicious activity is a critical challenge. This paper presents a novel approach for the identification of malicious nodes in WSNs by integrating directed acyclic graphs (DAGs) with the RC4 encryption algorithm. DAGs are employed to establish a hierarchical structure that enables efficient data flow and tracking of communication patterns across the network. By utilizing DAGs, the system can monitor the consistency and integrity of data transmission, making it easier to detect anomalies caused by malicious nodes. The RC4 encryption algorithm further strengthens the approach by securing the communication between nodes, preventing unauthorized access and tampering. In combination, DAGs and RC4 provide a robust framework for both detecting malicious nodes and securing data exchanges. Experimental simulations demonstrate that the proposed approach enhances network security by identifying compromised nodes with high accuracy while maintaining efficient communication and low overhead. This method offers a scalable and secure solution for protecting WSNs from malicious threats.

Effective gorilla troops optimization-based hierarchical clustering with HOP field neural network for intrusion detection

S. Hemalatha — 2024-10-16

An intrusion detection system (IDS) armed with signature and attack pattern databases as reference tools are used to protect computer networks from intrusion. This article provides a hybrid machine learning algorithm for gorilla troops optimization (GTO) integrating hierarchical clustering and Hopfield neural network. In this paper, the authors present a model to improve intrusion detection accuracy and contain high operational flexibility of these techniques. It is inspired by social behavior in gorillas and optimizes the clustering process HNN. Experimental results show that the proposed approach enhances the traditional methods in intrusion detection for a variety of intrusions and it presents an effective solution that can help cybersecurity application development better.

Chaotic-based optimization, based feature selection with shallow neural network technique for effective identification of intrusion detection

S. Hemalatha — 2024-10-16

The work in this paper attempts to deal with intrusion detection using a chaotic-based optimization technique and feature selection + shallow neural networks. The idea of chaotic systems is used to get randomness in the feature selection process, which can enable a shallow neural network to perform better for intrusion detection. Experiments on benchmark datasets reveal the effectiveness of this proposed solution by significant improvements in detection accuracy, false positive reduction at run-time and computational efficiency as compared to conventional methods.

Funding agencies in Tamil Nadu State Universities: A scientometric perspective

U. Perachiselvi — 2024-10-16

Scientometrics is a significant area of information science because it offers a unique variety of tools and techniques for managing knowledge in social and organizational contexts as well as for maintaining and evaluating information resources. The multiple scientometric components of the articles published by Tamil Nadu’s top eight universities between 1989 and 2023 were examined in this study. Data analysis shows that the average growth rate is increasing at a 9.76% annual pace. This extensive study finds the subtle interactions between funding agencies and research productivity in Tamil Nadu’s state universities. A comprehensive assessment of the funding landscape is carried out by methodically gathering data from eight notable universities, including Alagappa University, Anna University, Annamalai University, and others. With a remarkable total of 19,524 funding agencies discovered across various roles within these universities, the breadth of support highlights the importance of external money in propelling research efforts. Notably, Anna University appears as a frontrunner, with 4,162 funding agencies, demonstrating a strong network of support. The findings highlight the relevance of understanding how funding agencies affect research ecosystems within academic institutions.

Optimizing predictive accuracy: A comparative study of feature selection strategies in the healthcare domain

M. A. Shanthi — 2024-10-16

Feature selection is a critical preprocessing step in the development of machine learning models, particularly in the healthcare domain, where datasets often contain numerous features that may not contribute significantly to predictive performance. This study presents a comparative analysis of various feature selection techniques applied to healthcare datasets, evaluating their effectiveness in improving model accuracy and reducing computational costs. We investigate traditional filter-based methods, such as information gain and chi-square, alongside wrapper-based approaches and hybrid techniques that combine the strengths of both. Using multiple healthcare datasets encompassing diverse medical conditions, we assess the impact of these techniques on classification performance using metrics such as accuracy, precision, recall, and F1-score. Additionally, we analyze the robustness and scalability of each method in handling high-dimensional data. The findings reveal significant differences in performance, highlighting the strengths and weaknesses of each feature selection approach within the healthcare context. This comparative study provides valuable insights for researchers and practitioners, guiding them in selecting appropriate feature selection techniques to enhance predictive modeling in healthcare applications.

Smart grid precision: Evaluating machine learning models for forecasting of energy consumption from a smart grid

B. S. E. Zoraida — 2024-10-16

The widespread adoption of smart home technologies has led to a significant increase in the generation of high-frequency energy consumption data from smart grids. Accurate forecasting of energy consumption in smart homes is crucial for optimizing resource utilization and promoting energy efficiency. This research work investigates the precision of energy consumption forecasting within a smart grid environment, employing machine learning algorithms such as convolutional neural network (CNN), recurrent neural network (RNN), long short-term memory (LSTM), temporal fusion transformer (TFT) and Prophet. The CNN model extracts spatial features, while RNN and LSTM capture temporal dependencies in time series data. Prophet, recognized for handling seasonality and holidays, is included for comparative analysis. Utilizing a dataset from Pecan Street, Texas, performance metrics like mean absolute error (MAE), root mean squared error (RMSE), and mean absolute percentage error (MAPE) assess each model’s accuracy. This work aids in improving energy management systems, contributing to sustainable and efficient energy use in residential environments.

An efficient key establishment for pervasive healthcare monitoring

Sruthy M.S — 2024-10-16

Health is one of the issues that present more challenges in the world. These challenges not only come from the requirements of the region itself yet in addition result from outside conditions that impact individuals’ health conditions and access to therapeutic administrations. To augment the security strength of real-time healthcare applications in the IoT environment, a novel framework, namely, an Enhanced and IoT-based medical healthcare security scheme (EIMSS), has been proposed in this chapter. The proposed EIMSS adapts the AUP authentication technique proposed in the previous chapter for authentication while transferring the patient’s data. The proposed EIMSS approach offers flexible services to aged people like confidentiality, integrity, and authentication for protecting their vital biological and medical data. The simulation results, analysis and comparison confirm that the proposed EIMSS outperforms existing protocols with improved security strength.

A robust finger detection based sign language recognition using pattern recognition techniques

Subin M. Varghese — 2024-10-16

Sign language recognition based on finger detection is arguably the main sign language used by most dumb people. It has its own phonetics, grammar and syntax that set it apart from other sign languages. Research related to sign language (SL) is only now becoming standardized. Considering the challenge of recognizing SL, in this work a new method for recognizing SL dynamic gestures is proposed. Sign language (SL) translation systems can be used to help dumb people interact with normal people with the help of a computer. Most studies on continuous recognition of sign language are done by processing frames obtained from videos at regular/equal intervals. If a developed system is powerful enough to handle both static and dynamic motions, then it will be the best system for processing frames obtained from processing consecutive gestures. The algorithm developed for the gesture recognition system in SL formulates a vision-based approach using two-dimensional discrete sinusoidal transforms (DSTs) for image compression and self-organizing maps (SOMs), or self-organizing feature maps. Kohonen’s (SOFM) Neural Networks for Pattern Recognition, simulated in MATLAB. The system showed an accuracy rate of 91 percent.

An enhanced support vector machine bbased multiclass classification method for crop prediction

A. Tamilmani — 2024-10-16

Crop type classification is a fundamental task in precision agriculture, enabling informed decision-making for crop management and resource allocation. Support vector machines (SVMs) have emerged as robust and effective tools for multiclass classification tasks. This study explores the application of SVM-based multiclass classification techniques to accurately categorize various crop types based on remote sensing data. The SVM algorithm is employed to create decision boundaries that maximize the margin between different crop classes while minimizing classification errors. To enhance classification performance, various kernel functions such as linear, polynomial, and radial basis functions are evaluated to capture complex relationships within the data. The proposed SVM-based approach is compared with other commonly used classification methods to assess its superiority in terms of accuracy, precision, recall, and F1 score.

A literature review on the information literacy competency among scholars of co-education colleges and women’s colleges

E. J. David Prabahar — 2024-10-16

This literature review investigates information literacy competency (ILC) among scholars in coeducational and women’s colleges, exploring the disparities, influencing factors, and educational impacts within these distinct academic settings. Information literacy, a critical skill for navigating the digital era, empowers students to evaluate, access, and utilize information effectively. This review synthesizes findings from diverse studies, comparing ILC levels between scholars in coeducational and women’s colleges and considering variables such as academic discipline, institutional resources, and the pedagogical environment. Research indicates that while coeducational institutions provide a broader range of resources and peer interactions, women’s colleges often emphasize collaborative and inclusive pedagogies that may enhance ILC. However, disparities in competency levels persist due to variations in information literacy training and institutional support structures. This review identifies key areas where ILC training could be improved, particularly through targeted interventions tailored to the needs of each educational setting. The findings underscore the need for comprehensive information literacy programs to equip scholars with essential competencies, ultimately fostering academic success and lifelong learning across diverse educational contexts.

Fuzzy optimization trust aware clustering approach for the detection of malicious node in the wireless sensor networks

A. Rukmani — 2024-10-16

Wireless sensor networks (WSNs) are pivotal in various applications, ranging from environmental monitoring to military operations. However, their susceptibility to security threats, particularly from malicious nodes, poses significant challenges to network integrity and data reliability. This paper proposes an innovative methodology that integrates clustering with an optimization approach to effectively identify and mitigate malicious nodes in WSNs. In the proposed methodology, the network is divided into clusters, each managed by a cluster head responsible for monitoring the behavior of nodes within its cluster. Trust values are assigned to nodes based on parameters such as data forwarding accuracy, communication consistency, and energy consumption. These trust metrics are optimized using a sophisticated optimization algorithm, which fine-tunes the decision-making process for identifying malicious nodes. By leveraging clustering, the method efficiently distributes computational tasks, while the optimization algorithm enhances the accuracy of malicious node detection by dynamically adjusting trust thresholds. The approach not only reduces the incidence of false positives but also extends the network lifetime by preventing compromised nodes from disrupting network operations. This trust-aware, optimized clustering strategy offers a robust solution for securing WSNs in critical applications, ensuring reliable and secure data transmission across the network.

Exploring learning-assisted optimization for mobile crowd sensing

P. Ananthi — 2024-10-16

Introducing sensing mobile crowds (SMC), a novel paradigm for real-time location-dependent urban sensing data collection. It is critically important to optimize the SMC process such that it provides the highest sensing quality at the lowest feasible cost due to its practical use. As an alternative to the combinatorial optimization algorithms utilized in previous research, a new approach to SMC optimization is to apply learning approaches to extract knowledge, such as patterns in participants’ behavior or correlations in sensing data. In this work, we thoroughly research learning-assisted optimization approaches for SMC. Using the existing literature as a starting point, we will describe various learning and optimization methods and evaluate them from the perspectives of the task and the participant. How to combine different approaches to get a complete solution is also discussed. Lastly, we point out the limitations that exist at the moment, which might lead to research directions in the future.

The socio-technical opportunities and threats of crowdsensing

P. Ananthi — 2024-10-16

The exponential growth of mobile crowd sensing (MCS) has provided unparalleled opportunities to collect large-scale data through a network of mobile devices, empowering diverse applications in smart cities, healthcare, and environmental monitoring. However, the inherently participatory nature of MCS raises critical privacy concerns, as sensitive user information is often at risk of exposure. This literature review examines recent advancements in employing machine learning techniques to enhance privacy preservation in MCS frameworks. It explores methods such as federated learning, differential privacy, and encryption-enhanced neural networks that aim to minimize data leakage while maintaining model accuracy. Additionally, this review analyzes the efficacy and limitations of various privacy-preserving algorithms, particularly regarding their adaptability to different MCS contexts and their impact on computational overhead and communication efficiency. Through a comprehensive synthesis of current studies, this review highlights emerging trends, identifies research gaps, and suggests future directions for developing robust privacy-preserving machine learning models tailored to the unique demands of MCS systems.

Energy efficient routing with cluster approach in wireless networks – A literature review

Chinnadurai U — 2024-10-16

This literature review examines the cluster-based approaches in wireless networks, focusing on their effectiveness in conserving energy, enhancing routing efficiency, and improving network credibility. Wireless networks are increasingly employed in various applications, from internet of things (IoT) systems to mobile ad hoc networks, where energy efficiency is critical due to the limited battery life of devices. Cluster-based techniques group nodes into clusters to optimize resource utilization, facilitating energy conservation by enabling localized communication and reducing redundant transmissions. The review explores various clustering algorithms, such as low-energy adaptive clustering hierarchy (LEACH) and hybrid energy-efficient distributed clustering (HEED), highlighting their impact on network longevity and scalability. Additionally, the study addresses the challenges of maintaining robust routing protocols within clustered networks, emphasizing the importance of reliable data transmission and node credibility to mitigate risks from malicious attacks. By synthesizing current research findings, this review provides insights into the future directions of cluster-based strategies in wireless networks, suggesting potential enhancements to ensure efficient energy management and reliable network performance.

Application of data mining and machine learning approaches in the prediction of heart disease – A literature survey

S. Vanaja — 2024-10-16

Heart disease remains a leading cause of mortality worldwide, emphasizing the urgent need for effective classification and prediction methodologies. This literature review explores various data mining and machine learning approaches utilized in the classification and prediction of heart disease. We systematically analyze a diverse range of techniques, including decision trees, support vector machines, artificial neural networks, and ensemble methods, highlighting their strengths and limitations. The review further examines pre-processing methods, feature selection, and extraction techniques that significantly impact model performance. Additionally, we discuss the integration of hybrid approaches and deep learning methods, showcasing their potential to enhance predictive accuracy. Recent advancements in data handling and algorithmic efficiency are also highlighted, demonstrating the promising role of machine learning in addressing the complexities of heart disease diagnosis. This review aims to provide a comprehensive understanding of current trends and future directions in heart disease classification and prediction, paving the way for improved diagnostic tools and health outcomes.

Expanding the quantity of virtual machines utilized within an open-source cloud infrastructure

D. Jayadurga — 2024-10-16

As cloud computing continues to evolve, the efficient management and scalability of virtual machines (VMs) have become pivotal for maximizing performance and resource utilization, particularly within open-source cloud infrastructures. This literature review investigates existing approaches and methodologies focused on expanding the number of VMs in open-source cloud environments. Key topics include the impact of VM scaling on resource allocation, load balancing, and energy efficiency, as well as the role of orchestration tools and hypervisor optimization in handling large-scale VM deployments. Furthermore, the review assesses the challenges related to VM density, network latency, and system reliability alongside emerging strategies for enhancing VM elasticity through containerization, microservices, and distributed computing models. This study aims to provide a comprehensive understanding of current trends, innovations, and limitations in VM expansion, offering insights into the future of scalable virtual infrastructures in open-source cloud systems.

The role of big data in transforming human resource analytics: A literature review

Jayalakshmi K. — 2024-10-16

The growing integration of human resources (HR) analytics with big data analytics has transformed the way organizations manage their workforce, offering deeper insights into employee performance, recruitment strategies, talent retention, and overall organizational efficiency. This literature review explores the convergence of HR analytics and Big Data, highlighting key trends, methodologies, and challenges faced in this rapidly evolving field. It examines various HR analytics frameworks that leverage data-driven approaches to optimize human capital management, including predictive modeling, machine learning, and sentiment analysis. By reviewing a wide range of studies, this paper identifies the critical role of Big Data in enhancing decision-making processes, enabling more accurate predictions about employee behavior, and supporting strategic HR initiatives. Additionally, the review addresses ethical concerns related to data privacy, biases in algorithms, and the complexities of integrating Big Data tools into existing HR systems. The findings suggest that, while HR analytics has immense potential, the successful implementation of Big Data in HR requires a robust technological infrastructure, skilled professionals, and adherence to ethical standards. This review provides a comprehensive understanding of the current state of HR analytics and its future trajectory in the era of Big Data.

Advancements in image quality assessment: a comparative study of image processing and deep learning techniques

V. Karthikeyan — 2024-10-16

Image quality assessment (IQA) is a crucial field in image processing that ensures optimal performance in various applications such as medical imaging, surveillance, and multimedia systems. The evolution of IQA methods spans from traditional image processing techniques to the incorporation of advanced deep learning algorithms. This literature review aims to provide a comprehensive analysis of the methodologies used in image quality assessment, focusing on both full-reference, reduced-reference, and no-reference approaches. Traditional methods such as peak signal-to-noise ratio (PSNR) and structural similarity index (SSIM) are discussed alongside more recent deep learning-based approaches that leverage convolutional neural networks (CNNs), generative adversarial networks (GANs), and transformers for feature extraction and prediction. Deep learning models have demonstrated enhanced performance in complex tasks like noise reduction, image reconstruction, and compression artifacts correction. Additionally, this review highlights the challenges in IQA, including the subjectivity of human visual perception and the limitations of various algorithms in handling different types of distortions. It concludes by suggesting future research directions that integrate hybrid models combining classical techniques with deep learning to achieve more robust and efficient image quality evaluation.

Trust and security in wireless sensor networks: A literature review of approaches for malicious node detection

A. Rukmani — 2024-10-16

Wireless sensor networks (WSNs) are widely used in various fields such as environmental monitoring, healthcare, military applications, and smart cities. However, due to their decentralized nature, limited resources, and deployment in often hostile environments, WSNs are vulnerable to several security threats, especially malicious node attacks. Malicious nodes can disrupt network operations by dropping packets, injecting false data, or launching attacks such as sinkholes, wormholes, and Sybil attacks. To counter these threats, numerous malicious node detection methods have been proposed, ranging from trust-based models to anomaly detection techniques and clustering-based approaches. In parallel, the design of efficient routing protocols plays a critical role in ensuring energy-efficient and secure data transmission within the network. This review paper presents a comprehensive analysis of existing techniques for malicious node detection and efficient routing in WSNs