A Robust Deep Learning System for Motor Bearing Fault Detection: Leveraging Multiple Learning Strategies and a Novel Double Loss Function DOI Creative Commons
Khoa D. Tran,

Lam Pham,

Nguyễn Văn Anh

et al.

Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown

Published: Oct. 30, 2024

Abstract Motor bearing fault detection (MBFD) is vital for ensuring the reliability and efficiency of industrial machinery. Identifying faults early can prevent system breakdowns, reduce maintenance costs, minimize downtime. This paper presents an advanced MBFD using deep learning, integrating multiple training approaches: supervised, semi-supervised, unsupervised learning to improve classification accuracy. A novel double-loss function further enhances model’s performance by refining feature extraction from vibration signals. Our approach rigorously tested on well-known datasets: American Society Mechanical Failure Prevention Technology (MFPT), Case Western Reserve University Bearing Data Center (CWRU), Paderborn University's Condition Monitoring Damage in Electromechanical Drive Systems (PU). Results indicate that proposed method outperforms traditional machine models, achieving high accuracy across all datasets. These findings underline potential applying MBFD, providing a robust solution predictive settings supporting proactive management machinery health.

Language: Английский

Improving Prediction of Chronic Kidney Disease Using KNN Imputed SMOTE Features and TrioNet Model DOI Open Access
Nazik Alturki, Abdulaziz Altamimi, Muhammad Umer

et al.

Computer Modeling in Engineering & Sciences, Journal Year: 2024, Volume and Issue: 139(3), P. 3513 - 3534

Published: Jan. 1, 2024

Chronic kidney disease (CKD) is a major health concern today, requiring early and accurate diagnosis.Machine learning has emerged as powerful tool for detection, medical professionals are increasingly using ML classifier algorithms to identify CKD early.This study explores the application of advanced machine techniques on dataset obtained from University California, UC Irvine Machine Learning repository.The research introduces TrioNet, an ensemble model combining extreme gradient boosting, random forest, extra tree classifier, which excels in providing highly predictions CKD.Furthermore, K nearest neighbor (KNN) imputer utilized deal with missing values while synthetic minority oversampling (SMOTE) used class-imbalance problems.To ascertain efficacy proposed model, comprehensive comparative analysis conducted various models.The TrioNet KNN SMOTE outperformed other models 98.97% accuracy detecting CKD.This in-depth demonstrates model's capabilities underscores its potential valuable diagnosis CKD.

Language: Английский

Citations

10
Machine learning for classifying chronic kidney disease and predicting creatinine levels using at-home measurements DOI Creative Commons
Brady Metherall, Anna Berryman, Georgia S. Brennan

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Feb. 5, 2025

Abstract Chronic kidney disease (CKD) is a global health concern with early detection playing pivotal role in effective management. Machine learning models demonstrate promise CKD detection, yet the impact on and classification using different sets of clinical features remains under-explored. In this study, we focus creatinine prediction three features: at-home, monitoring, laboratory. We employ artificial neural networks (ANNs) random forests (RFs) dataset 400 patients 25 input features, which divide into feature sets. Using 10-fold cross-validation, calculate metrics such as accuracy, true positive rate (TPR), negative (TNR), mean squared error. Our results reveal RF achieves superior accuracy (92.5%) at-home over ANNs (82.9%). achieve higher TPR (92.0%), but lower TNR (67.9%) compared RFs (90.0% 95.8%, respectively). For monitoring laboratory both methods accuracies exceeding 98%. The R2 score for regression approximately 0.3 than features. Feature importance analysis identifies key variables hemoglobin blood urea, comorbidities hypertension diabetes mellitus, agreement previous studies. models, particularly RFs, exhibit diagnosis highlight significant detection. Moreover, may assist screening general population features—potentially increasing CKD, thus improving patient care offering hope more approach to managing prevalent condition.

Language: Английский

Citations

1
Exploring the Therapeutic Significance of microRNAs and lncRNAs in Kidney Diseases DOI Open Access
Luis Alberto Bravo-Vázquez, Sujay Paul,

Miriam Guadalupe Colín-Jurado

et al.

Genes, Journal Year: 2024, Volume and Issue: 15(1), P. 123 - 123

Published: Jan. 19, 2024

MicroRNAs (miRNAs) and long non-coding RNAs (lncRNAs) are two crucial classes of transcripts that belong to the major group (ncRNAs). These RNA molecules have significant influence over diverse molecular processes due their role as regulators gene expression. However, dysregulated expression these ncRNAs constitutes a fundamental factor in etiology progression wide variety multifaceted human diseases, including kidney diseases. In this context, past years, compelling evidence has shown miRNAs lncRNAs could be prospective targets for development next-generation drugs against diseases they participate number disease-associated processes, such podocyte nephron death, renal fibrosis, inflammation, transition from acute injury chronic disease, vascular changes, sepsis, pyroptosis, apoptosis. Hence, current review, we critically analyze recent findings concerning therapeutic inferences pathophysiological context Additionally, with aim driving advances formulation ncRNA-based tailored management discuss some key challenges future prospects should addressed forthcoming investigations.

Language: Английский

Citations

8
Advanced Predictive Analytics for Early Detection of Chronic Kidney Disease Using ML Models DOI
Divya Gopinath, Vasuki Rajaguru

Communications in computer and information science, Journal Year: 2025, Volume and Issue: unknown, P. 313 - 326

Published: Jan. 1, 2025

Language: Английский

Citations

0
ACD-ML: Advanced CKD detection using machine learning: A tri-phase ensemble and multi-layered stacking and blending approach DOI Creative Commons
Md Arif Hossain,

Shajreen Tabassum Diya,

Riasat Khan

et al.

Computer Methods and Programs in Biomedicine Update, Journal Year: 2025, Volume and Issue: unknown, P. 100173 - 100173

Published: Jan. 1, 2025

Language: Английский

Citations

0
A population based optimization of convolutional neural networks for chronic kidney disease prediction DOI Creative Commons

M. Priyadharshini,

V. Murugesh,

G.V. Samkumar

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: April 25, 2025

Language: Английский

Citations

0
Hybrid missing data imputation and novel weight convolution neural network classifier for chronic kidney disease diagnosis DOI Creative Commons

T. Saroja,

Y. Kalpana

Measurement Sensors, Journal Year: 2023, Volume and Issue: 27, P. 100715 - 100715

Published: Feb. 27, 2023

CKD (chronic kidney disease) have been identified as a serious public health concern globally. Machine learning models can successfully enable physicians to reach this aim because of their rapid and accurate identification performance. In paper, KNN (K Nearest Neighbor) imputations, which choose multiple full samples with the most comparable values for replacing missing utilized in work. Additionally, conventional MLT (Machine Learning Technique) need produce better outcomes than DLT (Deep technique). The prediction step may be sluggish, sensitive size data, filled irrelevant information when there is lot data. Tensor factorization ANFIS (Adaptive Neuro-Fuzzy Inference System) are added data imputations address issue. technique addressing feature selections called AWDBOA Weight Dynamic Butterfly Optimization Algorithm), inspired by nature. At same time, adjustable weights selection from dataset added. A modified form NN NWCNN (Novel Convolution Neural Network) classifier uses convolution rather standard matrix multiplication at least one its layers. layers hidden NWCNN, kernel functions improve or fine-tune classifier's parameters. was imperturbable UCI (University California, Irvine) ML repository had large number misplaced values. This work's proposed evaluated terms precision, recall, F1-score, sensitivitiy, specificity, accuracy 99.17%, 98.71%, 98.94%, 99.10%, 99.04% were obtained higher other models.

Language: Английский

Citations

4
A machine learning-based early diagnosis model for chronic kidney disease using SPegasos DOI
Monire Norouzi,

Elif Altıntaş Kahriman

Network Modeling Analysis in Health Informatics and Bioinformatics, Journal Year: 2024, Volume and Issue: 13(1)

Published: April 25, 2024

Language: Английский

Citations

1
RETRACTED ARTICLE: Optoelectronic sensor fault detection based predictive maintenance smart industry 4.0 using machine learning techniques DOI

Chenfeng Zhu,

Sihao Shao

Optical and Quantum Electronics, Journal Year: 2023, Volume and Issue: 55(13)

Published: Oct. 7, 2023

Language: Английский

Citations

2
Conceptual metaphor quantum correlation and radial basis extreme learning for predicting chronic kidney disease DOI
Muralidharan Jayashree,

A. N.

Computers & Electrical Engineering, Journal Year: 2024, Volume and Issue: 122, P. 109933 - 109933

Published: Dec. 2, 2024

Language: Английский

Citations

0