Optimal feature selection using novel flamingo search algorithm for classification of COVID-19 patients from clinical text DOI Creative Commons
Amir Yasseen Mahdi,

Siti Sophiayati Yuhaniz

Mathematical Biosciences & Engineering, Journal Year: 2023, Volume and Issue: 20(3), P. 5268 - 5297

Published: Jan. 1, 2023

Though several AI-based models have been established for COVID-19 diagnosis, the machine-based diagnostic gap is still ongoing, making further efforts to combat this epidemic imperative. So, we tried create a new feature selection (FS) method because of persistent need reliable system choose features and develop model predict virus from clinical texts. This study employs newly developed methodology inspired by flamingo's behavior find near-ideal subset accurate diagnosis patients. The best are selected using two-stage. In first stage, implemented term weighting technique, which that RTF-C-IEF, quantify significance extracted. second stage involves approach called improved binary flamingo search algorithm (IBFSA), chooses most important relevant proposed multi-strategy improvement process at heart improve algorithm. primary objective broaden algorithm's capabilities increasing diversity support exploring space. Additionally, mechanism was used performance traditional FSA make it appropriate FS issues. Two datasets, totaling 3053 1446 cases, were evaluate suggested based on Support Vector Machine (SVM) other classifiers. results showed IBFSA has compared numerous previous swarm algorithms. It noted, number subsets chosen also drastically reduced 88% obtained global optimal features.

Language: Английский

MMKE: Multi-trial vector-based monkey king evolution algorithm and its applications for engineering optimization problems DOI Creative Commons
Mohammad H. Nadimi-Shahraki, Shokooh Taghian, Hoda Zamani

et al.

PLoS ONE, Journal Year: 2023, Volume and Issue: 18(1), P. e0280006 - e0280006

Published: Jan. 3, 2023

Monkey king evolution (MKE) is a population-based differential evolutionary algorithm in which the single strategy and control parameter affect convergence balance between exploration exploitation. Since strategies have considerable impact on performance of algorithms, collaborating multiple can significantly enhance abilities algorithms. This our motivation to propose multi-trial vector-based monkey named MMKE. It introduces novel best-history trial vector producer (BTVP) random (RTVP) that effectively collaborate with canonical MKE (MKE-TVP) using approach tackle various real-world optimization problems diverse challenges. expected proposed MMKE improve global search capability, strike exploitation, prevent original from converging prematurely during process. The was assessed CEC 2018 test functions, results were compared eight metaheuristic As result experiments, it demonstrated capable producing competitive superior terms accuracy rate comparison comparative Additionally, Friedman used examine gained experimental statistically, proving Furthermore, four engineering design optimal power flow (OPF) problem for IEEE 30-bus system are optimized demonstrate MMKE's real applicability. showed handle difficulties associated able solve multi-objective OPF better solutions than

Language: Английский

Citations

44
A survey on binary metaheuristic algorithms and their engineering applications DOI Open Access
Jeng‐Shyang Pan, Pei Hu, Václav Snåšel

et al.

Artificial Intelligence Review, Journal Year: 2022, Volume and Issue: 56(7), P. 6101 - 6167

Published: Nov. 21, 2022

Language: Английский

Citations

65
Kidney Tumor Detection and Classification Based on Deep Learning Approaches: A New Dataset in CT Scans DOI Creative Commons

Dalia Alzu’bi,

Malak Abdullah,

Ismail Hmeidi

et al.

Journal of Healthcare Engineering, Journal Year: 2022, Volume and Issue: 2022, P. 1 - 22

Published: Oct. 22, 2022

Kidney tumor (KT) is one of the diseases that have affected our society and seventh most common in both men women worldwide. The early detection KT has significant benefits reducing death rates, producing preventive measures reduce effects, overcoming tumor. Compared to tedious time-consuming traditional diagnosis, automatic algorithms deep learning (DL) can save diagnosis time, improve test accuracy, costs, radiologist's workload. In this paper, we present models for diagnosing presence KTs computed tomography (CT) scans. Toward detecting classifying KT, proposed 2D-CNN models; three are concerning such as a 2D convolutional neural network with six layers (CNN-6), ResNet50 50 layers, VGG16 16 layers. last model classification four (CNN-4). addition, novel dataset from King Abdullah University Hospital (KAUH) been collected consists 8,400 images 120 adult patients who performed CT scans suspected kidney masses. was divided into 80% training set 20% testing set. accuracy results CNN-6 reached 97%, 96%, 60%, respectively. At same CNN-4 92%. Our achieved promising results; they enhance patient conditions high workload providing them tool automatically assess condition kidneys, risk misdiagnosis. Furthermore, increasing quality healthcare service change disease's track preserve patient's life.

Language: Английский

Citations

57
MTRRE-Net: A deep learning model for detection of breast cancer from histopathological images DOI
Soham Chattopadhyay, Arijit Dey, Pawan Kumar Singh

et al.

Computers in Biology and Medicine, Journal Year: 2022, Volume and Issue: 150, P. 106155 - 106155

Published: Sept. 30, 2022

Language: Английский

Citations

42
A Comprehensive Survey on Aquila Optimizer DOI Open Access
Buddhadev Sasmal, Abdelazim G. Hussien, Arunita Das

et al.

Archives of Computational Methods in Engineering, Journal Year: 2023, Volume and Issue: 30(7), P. 4449 - 4476

Published: June 7, 2023

Language: Английский

Citations

35
An efficient adaptive-mutated Coati optimization algorithm for feature selection and global optimization DOI Creative Commons
Fatma A. Hashim, Essam H. Houssein, Reham R. Mostafa

et al.

Alexandria Engineering Journal, Journal Year: 2023, Volume and Issue: 85, P. 29 - 48

Published: Nov. 17, 2023

The feature selection (FS) problem has occupied a great interest of scientists lately since the highly dimensional datasets might have many redundant and irrelevant features. FS aims to eliminate such features select most important ones that affect classification performance. Metaheuristic algorithms are best choice solve this combinatorial problem. Recent researchers invented adapted new algorithms, hybridized or enhanced existing by adding some operators In our paper, we added Coati optimization algorithm (CoatiOA). first operator is adaptive s-best mutation enhance balance between exploration exploitation. second directional rule opens way discover search space thoroughly. final enhancement controlling direction toward global best. We tested proposed mCoatiOA in solving) solving challenging problems from CEC'20 test suite. performance was compared with Dandelion Optimizer (DO), African vultures (AVOA), Artificial gorilla troops optimizer (GTO), whale (WOA), Fick's Law Algorithm (FLA), Particle swarm (PSO), Harris hawks (HHO), Tunicate (TSA). According average fitness, it can be observed method, mCoatiOA, performs better than other on 8 functions. It lower standard deviation values competitive algorithms. Wilcoxon showed results obtained significantly different those rival been as algorithm. Fifteen benchmark various types were collected UCI machine-learning repository. Different evaluation criteria used determine effectiveness method. achieved comparison published methods. mean 75% datasets.

Language: Английский

Citations

33
Discrete Improved Grey Wolf Optimizer for Community Detection DOI
Mohammad H. Nadimi-Shahraki,

Ebrahim Moeini,

Shokooh Taghian

et al.

Journal of Bionic Engineering, Journal Year: 2023, Volume and Issue: 20(5), P. 2331 - 2358

Published: May 18, 2023

Language: Английский

Citations

32
Memory-Based Sand Cat Swarm Optimization for Feature Selection in Medical Diagnosis DOI Open Access
Amjad Qtaish, Dheeb Albashish, Malik Braik

et al.

Electronics, Journal Year: 2023, Volume and Issue: 12(9), P. 2042 - 2042

Published: April 28, 2023

The rapid expansion of medical data poses numerous challenges for Machine Learning (ML) tasks due to their potential include excessive noisy, irrelevant, and redundant features. As a result, it is critical pick the most pertinent features classification task, which referred as Feature Selection (FS). Among FS approaches, wrapper methods are designed select appropriate subset In this study, two intelligent approaches implemented using new meta-heuristic algorithm called Sand Cat Swarm Optimizer (SCSO). First, binary version SCSO, known BSCSO, constructed by utilizing S-shaped transform function effectively manage nature in domain. However, BSCSO suffers from poor search strategy because has no internal memory maintain best location. Thus, will converge very quickly local optimum. Therefore, second proposed method devoted formulating an enhanced Binary Memory-based SCSO (BMSCSO). It integrated memory-based into position updating process exploit further preserve solutions. Twenty one benchmark disease datasets were used implement evaluate improved methods, BMSCSO. per results, BMSCSO acted better than terms fitness values, accuracy, number selected Based on obtained can efficiently explore feature domain optimal set.

Language: Английский

Citations

26
Bio-Inspired Feature Selection Algorithms With Their Applications: A Systematic Literature Review DOI Creative Commons
Tin H. Pham, Bijan Raahemi

IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 43733 - 43758

Published: Jan. 1, 2023

Based on the principles of biological evolution nature, bio-inspired algorithms are gaining popularity in developing robust techniques for optimization. Unlike gradient descent optimization methods, these metaheuristic computationally less expensive, and can also considerably perform well with nonlinear high-dimensional data. Objectives: To understand algorithms, application domains, effectiveness, challenges feature selection techniques. Method: A systematic literature review is conducted five major digital databases science engineering. Results: The primary search included 695 articles. After removing 263 duplicated articles, 432 studies remained to be screened. Among those, 317 irrelevant papers were removed. We then excluded 77 according exclusion criteria. Finally, 38 articles selected this study. Conclusion: Out studies, 28 discussed Swarm-based 2 studied Genetic Algorithms, 8 covered both categories. Considering 21 focused problems healthcare sector, while rest mainly investigated issues cybersecurity, text classification, image processing. Hybridization other BIAs was employed by approximately 18.5% papers, 13 out used S-shaped transfer functions. majority supervised classification methods such as k-NN SVM building fitness Accordingly, we conclude that future research should focus applying a diverse area applications finance social networks. And further exploration into enhancement quantum representation, rough set theory, chaotic maps, Lévy flight necessary. Additionally, suggest investigating functions besides S-shaped, V-shaped X-shaped. Moreover, clustering deep learning models constructing need further.

Language: Английский

Citations

24
Multiplayer battle game-inspired optimizer for complex optimization problems DOI
Yuefeng Xu, Rui Zhong, Chengqi Zhang

et al.

Cluster Computing, Journal Year: 2024, Volume and Issue: 27(6), P. 8307 - 8331

Published: April 10, 2024

Language: Английский

Citations

12