Procedia Computer Science最新文献

{"title":"Multi-faceted Review and Meta‑analysis to Predict Cardiac Outcomes using Machine Learning","authors":"Vinitha V ,&nbsp;Ranjima P ,&nbsp;Deepthika Karuppusamy ,&nbsp;Prajwalasimha S N ,&nbsp;Naveen Kulkarni ,&nbsp;Sharanabasappa Tadkal","doi":"10.1016/j.procs.2024.12.040","DOIUrl":"10.1016/j.procs.2024.12.040","url":null,"abstract":"<div><div>Acute cardiac arrest (CA) represents a significant challenge in medical prediction and intervention. Characterized by inadequate oxygen intake, CA often results in electrical imbalances within the heart, leading to irregular heartbeats and eventual cessation of blood flow. This condition can arise abruptly, necessitating prompt evaluation and intervention. Notably, cardiac arrest can occur without visible obstructions in the heart or brain, stemming instead from dysfunction in cardiac electrical activity and pumping efficiency. Individuals can be assessed for risk based on lifestyle, habits, and activities, although many may remain unaware of their susceptibility. Growing attention within the medical community reflects an increasing awareness of the need for early detection and preventive strategies. Clinicians utilize a variety of techniques to identify patients at risk of cardiac arrest, including electrocardiograms, Holter monitoring, echocardiography, and wearable technology. This paper reviews these methodologies and emphasizes the importance of early identification and risk assessment in improving outcomes for individuals at risk of acute cardiac arrest.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 394-403"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143377130","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Minimizing the Maximum Link Utilization for Traffic Engineering in SDN: A Comparative Analysis","authors":"U. Prabu ,&nbsp;V. Geetha","doi":"10.1016/j.procs.2024.12.032","DOIUrl":"10.1016/j.procs.2024.12.032","url":null,"abstract":"<div><div>Software Defined Network modernizes network architecture by supporting scalability, significant flexibility, and programmability. Its functioning has been extended to several areas such as cloud computing, edge computing, enterprise networks, data centers, and the Internet of Things (IoT). It also successfully enhances the functioning of traffic engineering. Traffic engineering involves optimization and dynamic control of traffic flows, minimizing congestion and maximum link utilization, centralized control, and network traffic management. It also enables the effective utilization of resources, boosts the user experience, and increases network performance. Minimizing maximum link utilization implies the distribution of network traffic across respective links or paths to use the offered resources of the network by preventing congestion. Its major goal is to guarantee that none of the individual links gets overloaded. Various approaches have been proposed in the past several years to address this problem. In this article, the state-of-the-art approaches are reviewed and analyzed. The problem formulation of each approach for minimizing the maximum link utilization is detailed with significant importance. The proposed algorithms, its type, and results are also depicted to have a better comparative analysis. Additionally, the type of graph considered while formulating the problem, topologies used to investigate the efficiency of the algorithm, performance metrics, and implementation platforms are listed to have a clear understanding of various approaches.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 296-305"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376665","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Federated Learning based Gender Classification in Heterogeneous and Distributed Data having Concept Drift","authors":"Vishwash Sharma ,&nbsp;VenkataHemant Kumar Reddy Challa ,&nbsp;Pasupuleti Pranavi ,&nbsp;Rimjhim Padam Singh","doi":"10.1016/j.procs.2024.12.033","DOIUrl":"10.1016/j.procs.2024.12.033","url":null,"abstract":"<div><div>Federated Learning is gaining significant traction in recent times, due to its ability to protect privacy while giving comparable results across various use cases. However, it often falls short in delivering good results in scenarios where the client devices at different places store huge and heterogeneous data. The studies explore the effectiveness of federated learning, when utilized for model training on heterogeneous data having concept drift nature and being placed on different client devices. The paper proposes a novel combination of EfficientNet-B0 model fine-tuned with adaptive and weighted normalization layers in a federated learning setup for gender classification in heterogeneous Fairface dataset having human images from different races. Several other Convolutional Neural Network (CNN) models namely, ResNet50, DenseNet121, GoogLeNet, AlexNet etc. have also been trained in different traditional and federated setups. Experiments revealed that EfficientNet-B0 outperformed the other models trained in all the scenarios with an outperforming accuracy of 88.18% in the traditional federated setup and 90.56% when fine-tuned using normalization methods and evaluated in a federated setup with four client devices. These findings highlight the potential of normalization techniques to improve the effectiveness of federated learning, especially when dealing with concept drift data.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 306-316"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376666","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Enhancing Mobile Application Security Through Android Threat Classification","authors":"Helen Josephine V L ,&nbsp;Manikandan Rajagopal ,&nbsp;Gayatri S ,&nbsp;Lakshmi K","doi":"10.1016/j.procs.2024.12.017","DOIUrl":"10.1016/j.procs.2024.12.017","url":null,"abstract":"<div><div>The Android application market has grown significantly, offering customers an ever-growing range of features to suit a variety of purposes. Users are exchanging more and more sensitive data thanks to the widespread usage of mobile applications, therefore safeguarding personal information is crucial. But this boom has also opened the door for a corresponding rise in cybersecurity risks, especially for malware and adware that target mobile devices. It is imperative to categorise mobile applications into distinct groups such as malware, adware, and benign in order to fortify the mobile ecosystem. This project’s primary objective is to create and apply cutting-edge machine learning algorithms that can precisely categorise mobile apps into groups including adware, malware, and benign apps. This will necessitate investigating various machine learning strategies and ensemble methods to improve classification accuracy and robustness. Multiple machine learning models were developed based on feature importance, utilizing various machine learning techniques. The evaluation metrics showcase the effectiveness of the final model, especially the Tuned XGBoost model. While achieving a high overall accuracy of 92.51%, the findings highlight the importance of considering diverse features beyond traditional flow-based ones, providing a more robust and complete perspective on mobile network security.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 154-164"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376669","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Neural Architecture Search-Driven Optimization of Deep Learning Models for Drug Response Prediction","authors":"Uday Kiran G ,&nbsp;Srilakshmi V ,&nbsp;Padmini G ,&nbsp;Sreenidhi G ,&nbsp;Venkata Ramana B ,&nbsp;Preetham Reddy G J","doi":"10.1016/j.procs.2024.12.019","DOIUrl":"10.1016/j.procs.2024.12.019","url":null,"abstract":"<div><div>In this study, the efficacy of various Neural Architecture Search (NAS) techniques for optimizing neural network architectures in drug response prediction is explored. Accurate prediction of drug responses is crucial for advancing personalized medicine, enabling personalized therapeutic interventions that enhance effectiveness and reduce adverse effects. Traditional models often rely on manually designed architectures, which may not fully capture the complex relationships among drug properties, genetic variations, and cellular phenotypes. An automated NAS approach is introduced to optimize neural network architectures for drug response prediction. The framework explores a defined search space using three techniques: Random Search, Q-Learning, and Bayesian Optimization. A modular architecture that integrates layers, activation functions, and dropout rates is proposed. Findings reveal the strengths and limitations of each NAS method, offering insights into effective model optimization strategies. Validation on publicly available pharmacogenomics datasets shows that NAS-optimized models outperform conventional deep learning and machine learning approaches, highlighting the potential of NAS to enhance predictive modelling in drug response and support personalized medicine and drug development.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 172-181"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376671","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Survey of Hybrid Deep Learning Autoencoders for Enhanced Visible Light Communication Systems","authors":"Vijayakumari K ,&nbsp;Anusudha K","doi":"10.1016/j.procs.2024.12.011","DOIUrl":"10.1016/j.procs.2024.12.011","url":null,"abstract":"<div><div>Visible Light Communication (VLC) systems, known for their high-speed data transmission and resistance to electromagnetic interference, face challenges like LED non-linearity, multipath distortion, and noise-induced signal degradation. To address these issues, deep learning approaches, particularly Autoencoders, have gained attention for enhancing VLC performance and reliability. This survey focuses on Hybrid Deep Learning Autoencoders (H-DLAEs) personalized for VLC systems. An overview of VLC technology is provided, highlighting its advantages and inherent challenges. It then explores the role of Deep Learning, emphasizing Autoencoders and their application in Communication Systems. The survey classifies existing H-DLAE models based on their architecture, including the integration of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) to capture spatial and temporal features of VLC signals. Performance metrics such as Bit Error Rate (BER), data throughput, and adaptability to varying channel conditions are used to evaluate these Hybrid Autoencoders. This study also discusses practical implementations of H-DLAE in real-world VLC scenarios, analyzing their benefits and limitations. Finally, significant research gaps are identified, and future directions are suggested, emphasizing the role of hybrid deep learning models in advancing VLC systems. This survey aims to provide a comprehensive analysis of H-DLAE for VLC, serving as a valuable resource for researchers and practitioners at the intersection of Deep Learning and Optical Wireless Communication.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 100-107"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376728","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Leveraging Deep Learning for Comprehensive Multilingual Hate Speech Detection","authors":"Atul Kumar Srivastava ,&nbsp;Mitali Srivastava ,&nbsp;Sanchali Das ,&nbsp;Vikas Jain ,&nbsp;Tej Bahadur Chandra","doi":"10.1016/j.procs.2025.01.044","DOIUrl":"10.1016/j.procs.2025.01.044","url":null,"abstract":"<div><div>Multilingual hate speech detection is a developing field of investigation that concentrates on the challenge of recognizing harmful content in various languages. With the explosive growth of social media platforms and the worldwide character of cyber communication, identifying hate speech in many language situations has become crucial. Existing detection models often struggle with language-specific nuances, cultural differences, and limited resources for less commonly spoken languages. This article conducts a widespread investigation of multilingual hate speech across 11 languages sourced from various datasets. By utilizing methods such as natural language processing (NLP), machine learning, and deep learning, researchers aim to create models that can effectively generalize across different languages. In low-resource scenarios, simpler models like LASER embeddings combined with logistic regression outperform ELMo-based models in high-resource contexts. The main objective of this research work lies in demonstrating promising results across a range of languages by integrating deep learning algorithms with multilingual pre-trained language models such as LASER and ELMo.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 832-840"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143376765","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimization-based Techniques Prediction Model in Determining Employee Turnover","authors":"James Cloyd M. Bustillo","doi":"10.1016/j.procs.2025.01.003","DOIUrl":"10.1016/j.procs.2025.01.003","url":null,"abstract":"<div><div>Employee turnover is a vital challenge for organizations, often resulting in high financial and operational costs. Traditional methods of turnover prediction have been limited in their ability to capture the complexity of factors influencing employee turnover, leading to ineffective retention strategies. This study proposes a novel approach by integrating Genetic Algorithms (GAs) with machine learning models to enhance the accuracy of turnover predictions. The research introduces an optimization-based prediction model utilizing GAs combined with Naïve Bayes and KNN algorithms. Four crossover operators, including traditional AX and three advanced operators, CAX, IBAX, and FMSAX, are employed to optimize the model’s feature selection and parameters. A dataset of employee turnover records from 2015 to 2023, encompassing various demographic and job-related factors, was used to train and test the model. The results demonstrate that the FMSAX operator, paired with KNN, achieves the highest predictive accuracy, precision, F-measure, and recall compared to other combinations. The findings suggest that the GA-optimized model significantly improves the ability to predict employee turnover and offers actionable insights for developing effective retention strategies. This study contributes to the field by providing a robust tool for human resource management, capable of adapting to dynamic workforce changes and minimizing turnover rates, ultimately enhancing organizational performance. The proposed model improves predictive accuracy and ensures its applicability across different organizational contexts by allowing the integration of new HR data.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 440-449"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143377147","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Simulative investigation of network scalability over MANET routing protocols","authors":"Satveer Kour ,&nbsp;Manjit Singh ,&nbsp;Varinder Kaur Attri ,&nbsp;Shashi Bhushan Rana ,&nbsp;Himali Sarangal ,&nbsp;Butta Singh","doi":"10.1016/j.procs.2025.01.009","DOIUrl":"10.1016/j.procs.2025.01.009","url":null,"abstract":"<div><div>The important feature that differentiates Mobile Ad-hoc NETworks (MANETs) to other Wireless or Wired Networks is the density of the nodes and mobility. MANET is a Network without Infrastructure, in which each node acts as a data sender, sink and router. Node density and Mobility are the key features that differentiate the MANET from the other wired and wireless networks. The self-configuring infrastructures without a network of mobile devices are connected via wireless connections. Any device on the MANET network can move to any direction and it frequently changes its connections with other devices. This network is an association of wireless nodes which may be dynamically configured at any time and from any location without utilizing the network infrastructure that is already in place. In multi-hop routing mode, nodes within radio range use other nodes as relays while distant nodes communicate directly over wireless links. The density of nodes or network scalability implies the quantity of nodes in the designed network scenario. The routing protocols are therefore designed for adapting the dynamic topology change while maximize its packet delivery ratio while minimize latency and throughput, with minimal number of packet losses and average jitter. Many researchers have been done the research work on MANET Routing Protocols. In this article, four routing protocols are used for the simulated scalable ad-hoc MANET implementations. The scalability refers to the variation in the number of nodes in a designed network scenario. For this research work, four scalable networks (25, 50, 75 100) are discussed. The five performance metrics are analyzed over four categories of node densities. The experimental results show that Ad Hoc Demand Vector (AODV) is 13.65% performed well in throughput than the other protocols; Dynamic Source Routing (DSR) is 12.66% performed well in goodput; DSR is 65% performed better than other protocols in the packet received; DSR is 10.3% performed better than other protocols in the received rate.</div></div>","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"252 ","pages":"Pages 500-509"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143377153","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Circular Supply Chains and Industry 4.0: An Analysis of Interfaces in Brazilian Foodtechs","authors":"Tiago Hilário Hennemann da Silva, Simone Sehnem","doi":"10.1016/j.procs.2024.01.134","DOIUrl":"https://doi.org/10.1016/j.procs.2024.01.134","url":null,"abstract":"","PeriodicalId":20465,"journal":{"name":"Procedia Computer Science","volume":"5 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2024-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"140666868","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}