Hybrid CNN and XGBoost Model Tuned by Modified Arithmetic Optimization Algorithm for COVID-19 Early Diagnostics from X-ray Images

Electronics, Journal Year: 2022, Volume and Issue: 11(22), P. 3798 - 3798

Published: Nov. 18, 2022

Developing countries have had numerous obstacles in diagnosing the COVID-19 worldwide pandemic since its emergence. One of most important ways to control spread this disease begins with early detection, which allows that isolation and treatment could perhaps be started. According recent results, chest X-ray scans provide information about onset infection, may evaluated so diagnosis can begin sooner. This is where artificial intelligence collides skilled clinicians’ diagnostic abilities. The suggested study’s goal make a contribution battling epidemic by using simple convolutional neural network (CNN) model construct an automated image analysis framework for recognizing afflicted data. To improve classification accuracy, fully connected layers CNN were replaced efficient extreme gradient boosting (XGBoost) classifier, used categorize extracted features layers. Additionally, hybrid version arithmetic optimization algorithm (AOA), also developed facilitate proposed research, tune XGBoost hyperparameters images. Reported experimental data showed approach outperforms other state-of-the-art methods, including cutting-edge metaheuristics algorithms, tested same framework. For validation purposes, balanced images dataset 12,000 observations, belonging normal, viral pneumonia classes, was used. method, tuned introduced AOA, superior performance, achieving accuracy approximately 99.39% weighted average precision, recall F1-score 0.993889, 0.993887 0.993887, respectively.

Language: Английский

---

Multi-Step Crude Oil Price Prediction Based on LSTM Approach Tuned by Salp Swarm Algorithm with Disputation Operator

Sustainability, Journal Year: 2022, Volume and Issue: 14(21), P. 14616 - 14616

Published: Nov. 7, 2022

The economic model derived from the supply and demand of crude oil prices is a significant component that measures development sustainability. Therefore, it essential to mitigate price volatility risks by establishing models will effectively predict prices. A promising approach application long short-term memory artificial neural networks for time-series forecasting. However, their ability tackle complex time series limited. decomposition-forecasting taken. Furthermore, machine learning accuracy highly dependent on hyper-parameter settings. in this paper, modified version salp swarm algorithm tasked with determining satisfying parameters improve performance prediction algorithm. proposed validated real-world West Texas Intermediate (WTI) data throughout two types experiments, one original decomposed after applying variation mode decomposition. In both cases, were adjusted conduct one, three, five-steps ahead predictions. According findings comparative analysis contemporary metaheuristics, was concluded hybrid forecasting, outscoring all competitors.

Language: Английский

---

On the Benefits of Using Metaheuristics in the Hyperparameter Tuning of Deep Learning Models for Energy Load Forecasting

Energies, Journal Year: 2023, Volume and Issue: 16(3), P. 1434 - 1434

Published: Feb. 1, 2023

An effective energy oversight represents a major concern throughout the world, and problem has become even more stringent recently. The prediction of load consumption depends on various factors such as temperature, plugged load, etc. machine learning deep (DL) approaches developed in last decade provide very high level accuracy for types applications, including time-series forecasting. Accordingly, number models this task is continuously growing. current study does not only overview most recent relevant DL supply demand, but it also emphasizes fact that many methods use parameter tuning enhancing results. To fill abovementioned gap, research conducted purpose manuscript, canonical straightforward long short-term memory (LSTM) model electricity tuned multivariate One open dataset from Europe used benchmark, performance LSTM one-step-ahead evaluated. Reported results can be benchmark hybrid LSTM-optimization forecasting power systems. work highlights leads to better when using metaheuristics all cases: while grid search achieves coefficient determination (R2) 0.9136, metaheuristic led worst result still notably with corresponding score 0.9515.

Language: Английский

---

Metaheuristic-Based Hyperparameter Tuning for Recurrent Deep Learning: Application to the Prediction of Solar Energy Generation

Axioms, Journal Year: 2023, Volume and Issue: 12(3), P. 266 - 266

Published: March 4, 2023

As solar energy generation has become more and important for the economies of numerous countries in last couple decades, it is highly to build accurate models forecasting amount green that will be produced. Numerous recurrent deep learning approaches, mainly based on long short-term memory (LSTM), are proposed dealing with such problems, but most may differ from one test case another respect architecture hyperparameters. In current study, use an LSTM a bidirectional (BiLSTM) data collection that, besides time series values denoting generation, also comprises corresponding information about weather. The research additionally endows hyperparameter tuning by means enhanced version recently metaheuristic, reptile search algorithm (RSA). output tuned neural network compared ones several other state-of-the-art metaheuristic optimization approaches applied same task, using experimental setup, obtained results indicate approach as better alternative. Moreover, best model achieved R2 0.604, normalized MSE value 0.014, which yields improvement around 13% over traditional machine models.

Language: Английский

---

The Explainable Potential of Coupling Metaheuristics-Optimized-XGBoost and SHAP in Revealing VOCs’ Environmental Fate

Atmosphere, Journal Year: 2023, Volume and Issue: 14(1), P. 109 - 109

Published: Jan. 4, 2023

In this paper, we explore the computational capabilities of advanced modeling tools to reveal factors that shape observed benzene levels and behavior under different environmental conditions. The research was based on two-year hourly data concentrations inorganic gaseous pollutants, particulate matter, benzene, toluene, m, p-xylenes, total nonmethane hydrocarbons, meteorological parameters obtained from Global Data Assimilation System. order determine model will be capable achieving a superior level performance, eight metaheuristics algorithms were tested for eXtreme Gradient Boosting optimization, while relative SHapley Additive exPlanations values used estimate importance each pollutant parameter prediction concentrations. According results, are mostly shaped by toluene finest aerosol fraction concentrations, in environment governed temperature, volumetric soil moisture content, momentum flux direction, as well hydrocarbons nitrogen oxide. types conditions which provided impact aerosol, temperature dynamics distinguished described.

Language: Английский

---

Optimizing long-short-term memory models via metaheuristics for decomposition aided wind energy generation forecasting

Artificial Intelligence Review, Journal Year: 2024, Volume and Issue: 57(3)

Published: Feb. 12, 2024

Abstract Power supply from renewable energy is an important part of modern power grids. Robust methods for predicting production are required to balance and demand avoid losses. This study proposed approach that incorporates signal decomposition techniques with Long Short-Term Memory (LSTM) neural networks tuned via a modified metaheuristic algorithm used wind generation forecasting. LSTM perform notably well when addressing time-series prediction, further hyperparameter tuning by version the reptile search (RSA) can help improve performance. The RSA was first evaluated against standard CEC2019 benchmark instances before being applied practical challenge. model has been tested two datasets hourly resolutions. predictions were executed without one, two, three steps ahead. Simulation outcomes have compared other cutting-edge metaheuristics. It observed introduced methodology exceed contenders, as later confirmed statistical analysis. Finally, this also provides interpretations best-performing models on both datasets, accompanied analysis importance impact each feature predictions.

Language: Английский

---

Tuning attention based long-short term memory neural networks for Parkinson’s disease detection using modified metaheuristics

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Feb. 21, 2024

Parkinson's disease (PD) is a progressively debilitating neurodegenerative disorder that primarily affects the dopaminergic system in basal ganglia, impacting millions of individuals globally. The clinical manifestations include resting tremors, muscle rigidity, bradykinesia, and postural instability. Diagnosis relies mainly on evaluation, lacking reliable diagnostic tests being inherently imprecise subjective. Early detection PD crucial for initiating treatments that, while unable to cure chronic condition, can enhance life quality patients alleviate symptoms. This study explores potential utilizing long-short term memory neural networks (LSTM) with attention mechanisms detect based dual-task walking test data. Given performance significantly inductance by architecture training parameter choices, modified version recently introduced crayfish optimization algorithm (COA) proposed, specifically tailored requirements this investigation. proposed optimizer assessed publicly accessible real-world gait dataset, results demonstrate its promise, achieving an accuracy 87.4187 % best-constructed models.

Language: Английский

---

Forecasting bitcoin: Decomposition aided long short-term memory based time series modeling and its explanation with Shapley values

Knowledge-Based Systems, Journal Year: 2024, Volume and Issue: 299, P. 112026 - 112026

Published: June 6, 2024

Bitcoin price volatility fascinates both researchers and investors, studying features that influence its movement. This paper expends on previous research examines time series data of various exogenous endogenous factors: Bitcoin, Ethereum, S&P 500, VIX closing prices; exchange rates the Euro GPB to USD; number Bitcoin-related tweets per day. A period three years (from September 2019 2022) is covered by dataset. two-layer framework introduced tasked with accurately forecasting price. In first layer, account for complexities in analyzed data, variational mode decomposition (VMD) extracts trends from series. second Long short-term memory hybrid Bidirectional long networks were used forecast prices several steps ahead. work also an enhanced variant sine cosine algorithm tune control parameters VMD neural attaining best possible performance. The main focus combining modified metaheuristics improve cryptocurrency value forecast. Two sets experiments conducted, without VMD. results have been contrasted models tuned seven other cutting-edge optimizers. Extensive experimental outcomes indicate can be forecasted great accuracy using selected decomposition. Additionally, model was analyzed, Shapley values indicated such as EUR/USD rates, Ethereum prices, GBP/USD a significant impact forecasts.

Language: Английский

---

A systematic review of AI-enhanced techniques in credit card fraud detection

Journal Of Big Data, Journal Year: 2025, Volume and Issue: 12(1)

Published: Jan. 14, 2025

Abstract The rapid increase of fraud attacks on banking systems, financial institutions, and even credit card holders demonstrate the high demand for enhanced detection (FD) systems these attacks. This paper provides a systematic review techniques using Artificial Intelligence (AI), machine learning (ML), deep (DL), meta-heuristic optimization (MHO) algorithms (CCFD). Carefully selected recent research papers have been investigated to examine effectiveness AI-integrated approaches in recognizing wide range These AI were evaluated compared discover advantages disadvantages each one, leading exploration existing limitations ML or DL-enhanced models. Discovering limitation is crucial future work robustness various key finding from this study demonstrates need continuous development models that could be alert latest fraudulent activities.

Language: Английский

---

Application of Natural Language Processing and Machine Learning Boosted with Swarm Intelligence for Spam Email Filtering

Mathematics, Journal Year: 2022, Volume and Issue: 10(22), P. 4173 - 4173

Published: Nov. 8, 2022

Spam represents a genuine irritation for email users, since it often disturbs them during their work or free time. Machine learning approaches are commonly utilized as the engine of spam detection solutions, they efficient and usually exhibit high degree classification accuracy. Nevertheless, sometimes happens that good messages labeled and, more often, some emails enter into inbox ones. This manuscript proposes novel approach by combining machine models with an enhanced sine cosine swarm intelligence algorithm to counter deficiencies existing techniques. The introduced was adopted training logistic regression tuning XGBoost part hybrid learning-metaheuristics framework. developed framework has been validated on two public high-dimensional benchmark datasets (CSDMC2010 TurkishEmail), extensive experiments conducted have shown model successfully deals high-degree data. comparative analysis other cutting-edge models, also based metaheuristics, proposed method obtains superior performance in terms accuracy, precision, recall, f1 score, relevant metrics. Additionally, empirically established superiority is using rigid statistical tests.

Language: Английский

---