Epilepsy Identification using Hybrid CoPrO-DCNN Classifier

International Journal of Computing and Digital Systems, Journal Year: 2024, Volume and Issue: 15(1), P. 783 - 796

Published: May 8, 2024

The Electroencephalogram (EEG) stands as a burgeoning frontier in the study of neuronal activity, offering rich tapestry information crucial for identifying abnormalities and addressing cognitive disorders irregularities.This paper delves into examination EEG from subjects exhibiting abnormalities, contrasting them with those normal subjects.Various topographical features such Mean, Entropy, Wavelet bands are meticulously evaluated compared.Inspired by adaptive hunting strategies observed coyotes, this introduces novel hybrid computational model that integrates deep learning architectures, aiming to amplify diagnostic accuracy.The methodology hinges upon development unique algorithm inspired intricate behaviors seamlessly fused potent data-driven capabilities neural networks.This is applied scrutinize data detection brain disorders, capitalizing on both biologically-inspired data-centric strengths learning.The results obtained innovative approach highly promising.The proposed scheme exhibits remarkable accuracy, achieving an impressive rate 98.65 per training (True Positive -TP) 98.82 utilizing k-fold validation.These preliminary findings underscore potential efficacy accurately discerning signals.However, it essential acknowledge these represent initial success form just fragment extensive evaluation process.This marks significant stride towards leveraging interdisciplinary insights, blending principles ethology advanced techniques tackle complex neurological challenges.By harnessing sophisticated nature alongside cutting-edge technological advancements, research endeavors carve path more nuanced precise tools understanding disorders.Further exploration refinement hold promise revolutionizing landscape neurodiagnostics, hope effective interventions treatments realm health.

Language: Английский

---

Tuberculosis detection bars on VGG19 transfer learning and Zebra Optimization Algorithm

EAI Endorsed Transactions on Pervasive Health and Technology, Journal Year: 2024, Volume and Issue: 10

Published: Aug. 22, 2024

Tuberculosis (TB) remains a significant global health challenge, necessitating accurate and efficient diagnostic tools. This study introduces novel approach combining VGG19, deep convolutional neural network model, with newly developed Zebra Optimization Algorithm (ZOA) to enhance the accuracy of TB detection from chest X-ray images. The Algorithm, inspired by social behavior zebras, was applied optimize hyperparameters VGG19 aiming improve model's generalizability performance. Our method evaluated using well-defined metric system that included accuracy, sensitivity, specificity. Results indicate combination ZOA significantly outperforms traditional methods, achieving high rate, which underscores potential hybrid approaches in image analysis.

Language: Английский

---

Enriched Deep Neural Network Improved by Chaotic Harris Hawk Optimizer for Prediction of Behavioural Traits of Individuals

Journal of Internet Services and Information Security, Journal Year: 2024, Volume and Issue: 14(4), P. 511 - 523

Published: Nov. 30, 2024

The enduring patterns of thoughts, feelings, and behaviors that set one person apart from another are referred to as personality traits. A identification system could help a corporation find hire suitable employees, enhance their business by understanding the preferences personalities clients, more. It necessitates prediction an individual’s classification determine behavioural traits using machine learning models. distribution class labels significantly affects training phase conventional ensemble algorithms resulting in overfitting problem affecting accuracy rate classification. Hence, this proposed work deep neural network with its dense layer understands pattern individuals based on questionnaire prepared demographics, education, employment attributes. However, parameters used assigned gradient descent method assigns random values. These values adjusted trial-and-error basis backpropagation method. This issue is solved improving performance adopting chaotic Harris hawk optimizer fine-tune hyperparameter DNN such weight, bias, layers DNN. prey searching behavior mapping balances both local global overcomes early convergence achieves highest compared other like models simulation results conducted 725 samples, 20 attributes for trait characteristics model Enriched Deep Neural Network improved Chaotic Hawk Optimizer Algorithm (EDNN-CHHOA) better 0.98% algorithms.

Language: Английский

---

Novel Multi-Classification Dynamic Detection Model for Android Malware Based on Improved Zebra Optimization Algorithm and LightGBM

Sensors, Journal Year: 2024, Volume and Issue: 24(18), P. 5975 - 5975

Published: Sept. 14, 2024

With the increasing popularity of Android smartphones, malware targeting platform is showing explosive growth. Currently, mainstream detection methods use static analysis to extract features software and apply machine learning algorithms for detection. However, can be less effective when faced with that employs sophisticated obfuscation techniques such as altering code structure. In order effectively detect improve accuracy, this paper proposes a dynamic model based on combination an Improved Zebra Optimization Algorithm (IZOA) Light Gradient Boosting Machine (LightGBM) model, called IZOA-LightGBM. By introducing elite opposition-based firefly perturbation strategies, IZOA enhances convergence speed search capability traditional zebra optimization algorithm. Then, employed optimize LightGBM hyperparameters multi-classification. The results from experiments indicate overall accuracy proposed IZOA-LightGBM CICMalDroid-2020, CCCS-CIC-AndMal-2020, CIC-AAGM-2017 datasets 99.75%, 98.86%, 97.95%, respectively, which are higher than other comparative models.

Language: Английский

---