A comprehensive review on federated learning based models for healthcare applications

Artificial Intelligence in Medicine, Journal Year: 2023, Volume and Issue: 146, P. 102691 - 102691

Published: Oct. 30, 2023

Language: Английский

---

Enhanced Cloud Storage Encryption Standard for Security in Distributed Environments

Electronics, Journal Year: 2023, Volume and Issue: 12(3), P. 714 - 714

Published: Feb. 1, 2023

With the growing number of cloud users, shared data auditing is becoming increasingly important. However, these schemes have issues with certificate management. Although there a certificate-shared scheme, it ineffective in dealing dynamic and protecting privacy. The verifier cannot access content to ensure integrity due security concerns. This paper proposes novel technique improve control. A enhanced storage retrieval mechanism used performance cloud’s mechanisms achieve this. evaluated concern upload, download, encryption, decryption time. As file size grows, so does time takes upload it. Similarly, taken encrypt files various formats sizes evidenced that depends on format. Thus, encryption increases as increases, demonstrating proposed system.

Language: Английский

---

Survey on Sensors and Smart Devices for IoT Enabled Intelligent Healthcare System

Wireless Personal Communications, Journal Year: 2023, Volume and Issue: 131(3), P. 1957 - 1995

Published: June 12, 2023

Language: Английский

---

Fault Prediction Recommender Model for IoT Enabled Sensors Based Workplace

Sustainability, Journal Year: 2023, Volume and Issue: 15(2), P. 1060 - 1060

Published: Jan. 6, 2023

Industry 5.0 benefits from advancements being made in the field of machine learning and Internet Things. Different sensors have been installed a variety IoT devices present different industries such as transportation, healthcare, manufacturing, agriculture, etc. The these should automatically predict errors due to extensive use urban living. To ensure integrity, precision, security, dependability fidelity sensor nodes, it is, therefore, necessary foresee faults before they occur. Additionally, more data is collected by every day, cloud computing becomes for sustainable proposed model emphasizes solution recommendations that occurred real-life smart mitigate at an early stage, which key requirement today’s offices. monitors real-time health through ML algorithm make efficient increase quality life. Through K-Nearest Neighbor, Decision Tree, Gaussian Naive Bayes Random Forest approach, fault prediction recommender has evaluated shows highest accuracy compared other classifiers. Several performance indicators recall, accuracy, F1 score precision were utilized examine model. results demonstrated effectiveness techniques applied predicting offices with observed best technique maximum 94.27%. In future, deep can also be bigger datasets provide accurate results.

Language: Английский

---

Prediction and classification of IoT sensor faults using hybrid deep learning model

Deleted Journal, Journal Year: 2024, Volume and Issue: 6(1)

Published: Jan. 20, 2024

Abstract The quality and reliability of internet thing (IoT) ecosystems heavily rely on accurate dependable sensor data. However, resource limited sensors are prone to failure due various factors like environmental disturbances electrical noise in which they can produce erroneous faulty measurements. These have significant consequences across different domains, including a threat safety critical systems. Though many researches been conducted, the existing literature primarily focuses fault detection data, while is useful, it still reactive approach that identifies faults after occurred, meaning actions taken has already impacted system, potentially leading negative consequences. In this study, proactive proposed by developing two-stage solution. first stage, hybrid convolutional neural network-long short term memory (CNN-LSTM) model was trained forecast measurements based historical second forecasted were passed network-multi layer perceptron (CNN-MLP) recognize types classify new accordingly. By passing values as input classification categorizing them normal, bias, drift, random or poly-drift, anticipated potential before manifest. publicly available Intel Lab data raw dataset used, annotated fault-injected. For regression, gated recurrent unit (GRU), Long (LSTM), bidirectional long (BiLSTM), network-gated (CNN-GRU), (CNN-LSTM), network-bidirectional (CNN-BiLSTM), evaluated compared their performance using root mean squared error (RMSE), (MSE) absolute (MAE) with 2-split time series cross-validation. CNN-LSTM outperformed other models Mean Absolute Error 2.0957 for 45 steps forecast. task, network (CNN), multi-layer (MLP), metrics accuracy, precision, recall, F1-score 5 tenfold cross-validations. CNN-MLP others accuracy 96.11% 99.33% 98.61% 98.81% poly-drift. average 4 98.21%, 0.3% increase from baseline work 97.91%. adopting prediction classification, research aims enhance efficiency IoT systems, allowing preventive measures be detrimental impact.

Language: Английский

---

Machine Learning Approaches for an Automatic Email Spam Detection

Published: April 21, 2023

With the rapid growth of internet users, spam emails have become a major problem. Spammers can easily create fake profiles and email accounts by pretending to be genuine people in sent emails. The spammers target who are unaware such scams. In today’s environment, is simple, quick, cost-effective way communicate but has various security threats which necessary identify maintain security. This situation necessitates having an inbuilt filtering system use effectively without being worried about losing personal details. goal this work discover predict early using classifiers. Machine learning methods provide most accurate classification. article contributes towards development detection model multiple classification tackle challenges helps technological progress privacy & employs technologies as naive bayes, K*, J48, random forest. Conclusively, when forest been used prediction classifier, output shown highest accuracy 95.48%.

Language: Английский

---

Integration of Machine Learning Algorithms with Cloud Computing for Real-Time Data Analysis

Journal for Research in Applied Sciences and Biotechnology, Journal Year: 2024, Volume and Issue: 3(2), P. 301 - 306

Published: May 8, 2024

As part of this study though, real-time data analysis is examined by exploring the combination machine learning algorithms with cloud computing. It does so defining optimization problems and solutions, as well outlining goals directions for development. The efficiency real time processing ability also amplified use an amalgamation computing though traditional systems are often associated high failure rates costs maintenance. Enhanced indexing, systematic control information storage retrieval, query main benefits obtainable from this. This because despite challenges such limited resources, integration has never been a problem even privacy. Peculiar trends Explainable AI, Automated ML, Continuous Intelligence present to substantially enhance operational proficiency decision-making.

Language: Английский

---

An effective technique to schedule priority aware tasks to offload data on edge and cloud servers

Measurement Sensors, Journal Year: 2023, Volume and Issue: 26, P. 100670 - 100670

Published: Jan. 10, 2023

Recent advancements in the Internet of Things (IoT) have enhanced quality life globally. Billions devices are brought under ambit IoT to make them smarter. IoT-based applications generating voluminous data and managing this widespread amount real-time through Cloud Technology, which offers high computational storage facilities. However, sending all cloud can bring serious concerns for applications, critical require instant action without any delay. Edge computing has recently emerged as an effective technology handle processing tasks locally. Additionally, important concern networks is response emergency on time increase performance large-scale systems. As such, scheduling becomes vital, where non-emergency be prioritized offload nearby edge servers respectively enhance Quality Service (QoS). The execution order allocating resources computation avoid delays two most factors that must addressed during task Computing. With aforementioned issues, we design a Priority aware Task Scheduling (PaTS) algorithm sensor schedule priority servers. problem formulated multi-objective function efficiency proposed evaluated using Bio-inspired NSGA-2 technique. overall improvement average queue delay, time, energy obtained 200 17.2%, 7.08% 11.4%, respectively. results show significant when compared with benchmark algorithms demonstrating effectiveness solution. Similarly, comparative increased from 1000 also shows subsequent improvements.

Language: Английский

---

AI-Driven Predictive Maintenance for Industrial Assets using Edge Computing and Machine Learning

Journal for Research in Applied Sciences and Biotechnology, Journal Year: 2024, Volume and Issue: 3(1), P. 363 - 367

Published: Feb. 28, 2024

The increasing complexity and scale of industrial assets, such as machinery, equipment, infrastructure, have led to a growing need for effective predictive maintenance strategies. Traditional time-based or reactive approaches often fall short in addressing the dynamic nature asset degradation failure patterns. This study explores integration artificial intelligence (AI) machine learning (ML) algorithms with edge computing develop an intelligent framework assets. By processing sensor data executing ML models closer source, at edge, this approach enables real-time anomaly detection, remaining useful life (RUL) estimation, proactive scheduling. paper outlines key methods involved, including preprocessing, feature engineering, model development, deployment on devices. It also discusses benefits integration, reduced downtime, improved reliability, enhanced operational efficiency. Furthermore, highlights emerging trends, transfer learning, ensemble modeling, adaptive which enhance flexibility, accuracy, adaptability AI-driven system. findings demonstrate transformative potential synergy, empowering operations transition from maintenance, ultimately optimizing performance reducing costs.

Language: Английский

---

An intelligent forecasting system in Internet of Agriculture Things sensor network

Ad Hoc Networks, Journal Year: 2025, Volume and Issue: unknown, P. 103752 - 103752

Published: Jan. 1, 2025

Language: Английский

---