A novel RFE-GRU model for diabetes classification using PIMA Indian dataset

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

Published: Jan. 6, 2025

Abstract Diabetes is a long-term condition characterized by elevated blood sugar levels. It can lead to variety of complex disorders such as stroke, renal failure, and heart attack. requires the most machine learning help diagnose diabetes illness at an early stage, it cannot be treated adds significant complications our health-care system. The PIMA Indian dataset (PIDD) was used for classification in several studies, includes 768 instances 9 features; eight features are predictors, one feature target. Firstly, we performed preprocessing stage that mean imputation data normalization. Afterwards, trained extracted using various types Machine Learning (ML); Random Forest (RF), Logistic Regression (LR), K-Nearest neighbor (KNN), Naïve Bayes (NB), Histogram Gradient Boost (HGB), Gated Recurrent Unit (GRU) models. To achieve PIDD, new model called Recursive Feature Elimination-GRU (RFE-GRU) proposed this paper. RFE vital selecting training important predicting target variable. While GRU handles challenge vanishing inflating gradient results from RFE. Several predictive evaluation metrics, including precision, recall, F1-score, accuracy, Area Under Curve (AUC) achieved 90.50%, 90.70%, 0.9278, respectively, verify validate execution RFE-GRU model. comparative showed better than other

Language: Английский

---

Hybrid Intelligent Control System for Adaptive Microgrid Optimization: Integration of Rule-Based Control and Deep Learning Techniques

Energies, Journal Year: 2024, Volume and Issue: 17(10), P. 2260 - 2260

Published: May 8, 2024

Microgrids (MGs) have evolved as critical components of modern energy distribution networks, providing increased dependability, efficiency, and sustainability. Effective control strategies are essential for optimizing MG operation maintaining stability in the face changing environmental load conditions. Traditional rule-based systems extensively used due to their interpretability simplicity. However, these frequently lack flexibility complex system dynamics. This paper provides a novel method called hybrid intelligent adaptive that integrates basic deep learning techniques, including gated recurrent units (GRUs), neural networks (RNNs), long short-term memory (LSTM). The main target this approach is improve management performance by combining strengths techniques. These techniques readily enhance adapt decisions based on historical data domain-specific rules, leading increasing stability, resilience MG. Our results show proposed optimizes operation, especially under demanding conditions such variable renewable supply unanticipated fluctuations. study investigates special RNN architectures hyperparameter optimization with aim predicting power consumption generation within system. promising highest-performing models indicating high accuracy efficiency prediction. finest-performing model accomplishes an R2 value close 1, representing strong correlation between predicted actual values. Specifically, best achieved 0.999809, MSE 0.000002, MAE 0.000831.

Language: Английский

---

Comparative Analysis of Currency Exchange and Stock Markets in BRICS Using Machine Learning to Forecast Optimal Trends for Data-Driven Decision Making

Journal of Economics Finance and Accounting Studies, Journal Year: 2025, Volume and Issue: 7(1), P. 26 - 48

Published: Jan. 8, 2025

The BRICS nations’ economies show that the countries are global financial powerhouses whose currency exchange rates and stock markets have influence globally. In this paper, analysis of forecast trends in both Currency Exchange Stock Markets using a dual layered machine learning approach exposing models such as Long Short Term Memory (LSTM), Random Forest, Gradient Boosting Support vector machines (SVM) is conducted. Their performance tested twice, first on then market data, to compare them basis predictive power deliver actionable insights. Each model applied separately, study mainly uses extensive historical datasets from economies. Benchmarking done metrics Mean Absolute Error (MAE), Root Square (RMSE) R-squared values. For exchange, LSTM turned out be most effective it can handle sequence time series data. best for forecasting was achieved by Boosting, which adept at finding complex nonlinear relationships. Forest proved consistent across Datasets but SVM found challenged Scalability Data Complexity, with relatively lower accuracy. research goes repeat comparative each different models, illustrate subtle differences between techniques their capacity effectively process all varieties. Predictive accuracy reliability further enhanced reconcile conflicting creating an ensemble algorithms. These findings provide robust framework informed decision making stakeholders identify more stable hence profitable context. results add expansion application finance demonstrating how tailored algorithms offer significant economic planning investment strategy plans.

Language: Английский

---

Data preprocessing techniques and neural networks for trended time series forecasting

Applied Soft Computing, Journal Year: 2025, Volume and Issue: unknown, P. 113063 - 113063

Published: March 1, 2025

Language: Английский

---

Correction to: Advanced series decomposition with a gated recurrent unit and graph convolutional neural network for non‑stationary data patterns

Journal of Cloud Computing Advances Systems and Applications, Journal Year: 2024, Volume and Issue: 13(1)

Published: March 27, 2024

Language: Английский

---

Unraveling climate trends in the mediterranean: a hybrid machine learning and statistical approach

Modeling Earth Systems and Environment, Journal Year: 2024, Volume and Issue: 10(5), P. 6255 - 6277

Published: Aug. 6, 2024

Abstract This study presents a comprehensive spatiotemporal analysis of sea surface temperatures (SST) and air (TAS) across 15 Mediterranean coastal stations, leveraging centennial-scale data to analyze regional climate dynamics. The modeling framework integrates three sequential phases: preprocessing, statistical analysis, advanced machine learning techniques, creating robust analytical pipeline. preprocessing phase harmonizes diverse datasets, addresses missing values, applies transformations ensure consistency. employs the Pettitt test for change point detection linear trend unveil underlying patterns. utilizes K-means clustering regime classification implements tailored Convolutional Neural Networks (CNNs) cluster-specific future anomaly projections. Results marked anthropogenic signal, with contemporary observations consistently surpassing historical baselines. Breakpoint analyses assessments reveal heterogeneous climatic shifts, pronounced warming in northern Mediterranean. Notably, Nice Ajaccio exhibit highest SST increases (0.0119 0.0113 °C/decade, respectively), contrasting more modest trends Alexandria (0.0052 °C/decade) Antalya (0.0047 eastern application CNN projections provides granular insights into differential trajectories. By 2050, cooler northwestern zones are projected experience dramatic anomalies approximately 3 °C above average, corresponding TAS 2.5 °C. In contrast, warmer southern regions display subdued patterns, 1.5–2.5 by mid-century. research’s importance is highlighted its potential inform adaptation strategies contribute theoretical understanding dynamics, advancing efforts.

Language: Английский

---

Critical analysis of Parkinson’s disease detection using EEG sub-bands and gated recurrent unit

Engineering Science and Technology an International Journal, Journal Year: 2024, Volume and Issue: 59, P. 101855 - 101855

Published: Oct. 18, 2024

Language: Английский

---