A two-phase cuckoo search based approach for gene selection and deep learning classification of cancer disease using gene expression data with a novel fitness function

Multimedia Tools and Applications, Journal Year: 2024, Volume and Issue: 83(28), P. 71721 - 71752

Published: Feb. 6, 2024

Language: Английский

---

Deep learning approach for brain tumor classification using metaheuristic optimization with gene expression data

International Journal of Imaging Systems and Technology, Journal Year: 2023, Volume and Issue: 34(2)

Published: Dec. 16, 2023

Abstract This study addresses the critical challenge of accurately classifying brain tumors using artificial intelligence. Early detection is crucial, as untreated can be fatal. Despite advances in AI, remains a challenging task. To address this challenge, we propose novel optimization approach called PSCS combined with deep learning for tumor classification. optimizes classification process by improving Particle Swarm Optimization (PSO) exploitation Cuckoo search (CS) algorithm. Next, classified gene expression data Deep Learning (DL) to identify different groups or classes related particular along technique. The proposed technique DL achieves much better accuracy than other existing and Machine models evaluation matrices such Recall, Precision, F1‐Score, confusion matrix. research contributes AI‐driven diagnosis classification, offering promising solution improved patient outcomes.

Language: Английский

---

Skin cancer classification based on an optimized convolutional neural network and multicriteria decision-making

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

Published: July 27, 2024

Abstract Skin cancer is a type of disease in which abnormal alterations skin characteristics can be detected. It treated if it detected early. Many artificial intelligence-based models have been developed for detection and classification. Considering the development numerous according to various scenarios selecting optimum model was rarely considered previous works. This study aimed develop classification select model. Convolutional neural networks (CNNs) form AlexNet, Inception V3, MobileNet V2, ResNet 50 were used feature extraction. Feature reduction carried out using two algorithms grey wolf optimizer (GWO) addition original features. images classified into four classes based on six machine learning (ML) classifiers. As result, 51 with different combinations CNN algorithms, without GWO ML To best results, multicriteria decision-making approach utilized rank alternatives by perimeter similarity (RAPS). Model training testing conducted International Imaging Collaboration (ISIC) 2017 dataset. Based nine evaluation metrics RAPS method, AlexNet algorithm classical yielded model, achieving accuracy 94.5%. work presents first benchmarking many models. not only reduces time spent but also improves accuracy. The method has proven its robustness problem

Language: Английский

---

ADVANCED GENETIC ALGORITHM (GA)-INDEPENDENT COMPONENT ANALYSIS (ICA) ENSEMBLE MODEL FOR PREDICTING TRAPPED HUMANS THROUGH HYBRID DIMENSIONALITY REDUCTION

Scientific African, Journal Year: 2025, Volume and Issue: unknown, P. e02564 - e02564

Published: Jan. 1, 2025

Language: Английский

---

Secretary bird optimization algorithm based on quantum computing and multiple strategies improvement for KELM diabetes classification

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

Published: Jan. 30, 2025

Abstract The classification of chronic diseases has long been a prominent research focus in the field public health, with widespread application machine learning algorithms. Diabetes is one high prevalence worldwide and considered disease its own right. Given nature this condition, numerous researchers are striving to develop robust algorithms for accurate classification. This study introduces revolutionary approach accurately classifying diabetes, aiming provide new methodologies. An improved Secretary Bird Optimization Algorithm (QHSBOA) proposed combination Kernel Extreme Learning Machine (KELM) diabetes prediction model. First, (SBOA) enhanced by integrating particle swarm optimization search mechanism, dynamic boundary adjustments based on optimal individuals, quantum computing-based t-distribution variations. performance QHSBOA validated using CEC2017 benchmark suite. Subsequently, used optimize kernel penalty parameter $$\:C$$ bandwidth $$\:c$$ KELM. Comparative experiments other models conducted datasets. experimental results indicate that QHSBOA-KELM model outperforms comparative four evaluation metrics: accuracy (ACC), Matthews correlation coefficient (MCC), sensitivity, specificity. offers an effective method early diagnosis diabetes.

Language: Английский

---

Risk prediction of hyperuricemia based on particle swarm fusion machine learning solely dependent on routine blood tests

BMC Medical Informatics and Decision Making, Journal Year: 2025, Volume and Issue: 25(1)

Published: March 14, 2025

Hyperuricemia has seen a continuous increase in incidence and trend towards younger patients recent years, posing serious threat to human health highlighting the urgency of using technological means for disease risk prediction. Existing prediction models hyperuricemia typically include two major categories indicators: routine blood tests biochemical tests. The potential alone not yet been explored. Therefore, this paper proposes model that integrates Particle Swarm Optimization (PSO) with machine learning, which can accurately assess by relying solely on data. In addition, an interpretability method based Explainable Artificial Intelligence(XAI) is introduced help medical staff understand how makes decisions. This uses Cohen's d value compare differences indicators between non-hyperuricemia identifies factors through multivariate logistic regression. Subsequently, constructed parameter optimization five learning PSO algorithm. accuracy sensitivity proposed particle swarm fusion Stacking reach 97.8% 97.6%, marking improvement over 11% compared state-of-the-art models. Finally, analysis affecting results conducted XAI method. also developed Health Portrait Platform models, enabling real-time online assessment. Since only test data are used, new better feasibility scalability, providing valuable reference assessing occurrence.

Language: Английский

---

Advanced machine learning framework for enhancing breast cancer diagnostics through transcriptomic profiling

Discover Oncology, Journal Year: 2025, Volume and Issue: 16(1)

Published: March 17, 2025

This study proposes an advanced machine learning (ML) framework for breast cancer diagnostics by integrating transcriptomic profiling with optimized feature selection and classification techniques. A dataset of 1759 samples (987 patients, 772 healthy controls) was analyzed using Recursive Feature Elimination, Boruta, ElasticNet selection. Dimensionality reduction techniques, including Non-Negative Matrix Factorization (NMF), Autoencoders, transformer-based embeddings (BioBERT, DNABERT), were applied to enhance model interpretability. Classifiers such as XGBoost, LightGBM, ensemble voting, Multi-Layer Perceptron, Stacking trained grid search cross-validation. Model evaluation conducted accuracy, AUC, MCC, Kappa Score, ROC, PR curves, external validation performed on independent 175 samples. XGBoost LightGBM achieved the highest test accuracies (0.91 0.90) AUC values (up 0.92), particularly NMF BioBERT. The Voting method exhibited best accuracy (0.92), confirming its robustness. Transformer-based techniques significantly improved performance compared conventional approaches like PCA Decision Trees. proposed ML enhances diagnostic interpretability, demonstrating strong generalizability dataset. These findings highlight potential precision oncology personalized diagnostics.

Language: Английский

---

ImagTIDS: an internet of things intrusion detection framework utilizing GADF imaging encoding and improved Transformer

Complex & Intelligent Systems, Journal Year: 2024, Volume and Issue: 11(1)

Published: Dec. 24, 2024

As the Internet of Things (IoT) technology becomes extensively deployed, IoT security issues are increasingly prominent. The traffic patterns complex and high-dimensional, which makes it difficult to distinguish tiny differences between normal malicious samples. To tackle above problems, we propose an intrusion detection architecture based on Gramian angular difference fields (GADF) imaging improved Transformer, named ImagTIDS. Firstly, encode network data into images using GADF preserve more robust temporal global features, then a model ImagTrans for extracting local features from images. ImagTIDS utilizes self-attention mechanism dynamically adjust attention weights adaptively focus important effectively suppressing adverse effects redundant features. Furthermore, due serious class imbalance problem in detection, utilize Focal Loss scale gradient reduce simple samples hard-to-classify classes. Finally, validate effectiveness proposed method publicly available datasets ToN_IoT DS2OS, experimental results show that achieves superior performance higher robustness compared other remarkable methods.

Language: Английский

---

Prediction of Crime Rate in Diverse Environs Using Hybrid Classifier

International Journal of Electronics and Communication Engineering, Journal Year: 2024, Volume and Issue: 11(2), P. 80 - 91

Published: Feb. 2, 2024

Crime is the fear and terror among populace worldwide. an inherent component of hazards we encounter daily. In recent times, mass media has extensively covered numerous criminal incidents, including theft, rape sexual offenses, robbery, murder, kidnappings. Various works have been produced to understand factors that lead individual committing a act, potential dangers involved, strategies prevent it. The crime computation technique aims forecast rates, enabling police officers avoid such incidents effectively. Based on this, novel prediction approach utilizing hybrid classifier suggested. An evaluation proposed method was conducted using several criteria. performance this recently constructed model evaluated by comparing it with established models as Genetic Algorithm, Particle Swarm Optimization, Firefly Algorithm. measures, error rate, sensitivity, specificity, precision, execution time, are used for comparison. results, most optimal compared other current models. suggested executed JAVA programming language.

Language: Английский

---

Ant-based feature and instance selection for multiclass imbalanced data

IEEE Access, Journal Year: 2024, Volume and Issue: 12, P. 133952 - 133968

Published: Jan. 1, 2024

Language: Английский

---