Machine Learning in Reducing E-Waste

Advances in computational intelligence and robotics book series, Journal Year: 2025, Volume and Issue: unknown, P. 105 - 126

Published: Feb. 27, 2025

E-waste is a waste which gathered from various sources, Household sector considered as biggest source of generation e-waste. consists components characterized hazardous and non-hazardous also contain approximately 1000 substances categorized in this category. comprising ferrous non-ferrous metals ceramics other items. When e-waste gets dismantled continuously processed it jeopardizes the health environment, surroundings. composition bio accumulative toxic containing like chromium, mercury, (Arora et al., 2024). Machine learning plays very important role regulating In context to urban segments, week by calculated through building prescient model, creating gradient boosting regression tree (GBRT) neutral network machine calculations. By incorporating learning, calculations will provide exact accurateness algorithm. A convolutional neural was created bifurcate into different countries. These categories are follows: cell phone, remote controller, battery, light bulb.

Language: Английский

---

TOWARDS SMARTER CYBER DEFENSE: LEVERAGING DEEP LEARNING FOR THREAT IDENTIFICATION AND PREVENTION

FUDMA Journal of Sciences, Journal Year: 2025, Volume and Issue: 9(3), P. 122 - 128

Published: March 31, 2025

The increasing sophistication of cyber threats has rendered traditional security measures inadequate, necessitating the adoption deep learning-based techniques for enhanced threat detection and prevention. This study develops a Sequential Neural Network (SNN) model to improve cybersecurity defenses by identifying malicious activities with greater accuracy. is trained on CERT Insider Threat v6.2 datasets, utilizing user activity modeling detect anomalous behavior effectively. Performance evaluation reveals that achieved an accuracy 67%, precision, recall, F1-score all at 0.67, indicating balanced but moderate classification capability. AUC-ROC score 0.67 further suggests while surpasses random classification, refinements are necessary practical deployment. confusion matrix analysis highlights challenges in distinguishing between certain threats, resulting misclassifications false positives. Despite these challenges, proposed learning approach demonstrates potential SNNs detecting complex attack patterns methods often fail recognize. However, issues such as class imbalance, interpretability, computational overhead must be addressed robustness. Future research will focus enhancing architectures, optimizing hyperparameters, integrating explainable AI reduce positive rates. By leveraging learning, this contributes development smarter more adaptive solutions, capable responding evolving real time.

Language: Английский

---

Determinants of adopting web-based systems for e-waste management and ensuring sustainable environment: Evidence from Bangladesh

Cleaner Waste Systems, Journal Year: 2025, Volume and Issue: unknown, P. 100282 - 100282

Published: April 1, 2025

Language: Английский

---

ENHANCED PREDICTION OF CORONARY ARTERY DISEASE USING LOGISTIC REGRESSION

FUDMA Journal of Sciences, Journal Year: 2025, Volume and Issue: 9(3), P. 201 - 208

Published: March 31, 2025

Coronary Artery Disease (CAD) remains a leading cause of global morbidity and mortality, emphasizing the urgent need for accurate interpretable prediction models to facilitate timely interventions improve patient outcomes. This study investigates application Logistic Regression CAD prediction, leveraging dataset 303 patients 13 clinical features obtained from UCI Machine Learning Repository. Recognizing limitations traditional risk assessment methods, this research explores potential enhance accuracy through streamlined easily implementable approach. The dataset, which encompasses demographic factors, measurements, lifestyle indicators, was subjected rigorous analysis evaluate model's performance. A model developed using Python's scikit-learn library assessed comprehensive set evaluation metrics, including accuracy, precision, recall, F1-score, Area Under Receiver Operating Characteristic curve (AUC-ROC). On test 61 instances, achieved an overall 82%, demonstrating its ability correctly classify individuals with without CAD. precision recall scores Class 0 (absence CAD) were 79% respectively, while 1 (presence CAD), 84% indicating balanced performance across both classes. exhibited AUC-ROC 0.89, signifying strong discriminatory ability. These findings suggest that can serve as valuable tool assessment, providing foundation more advanced predictive contributing improved cardiovascular health management...

Language: Английский

---

Deep Learning System for E-Waste Management

Published: Oct. 16, 2024

The deep learning system for e-waste management presented in this proposal is a transformative solution designed to address the escalating challenges of garbage collection and urban environments. Rapid urbanization has resulted increased waste generation, necessitating more intelligent efficient approach disposal. This integrates cutting-edge technologies, primarily Artificial Intelligence (AI), improve processes, enhance resource utilization, contribute creation cleaner sustainable spaces. Urban areas are experiencing unprecedented growth, leading surge volume generated daily; as such, traditional systems struggle cope with influx, resulting environmental pollution, compromised public health, inefficient utilization. proposed model accuracy 83% seeks revolutionize existing practices by leveraging capabilities AI. aim research develop sequential neural network using Keras TensorFlow image analysis: convolutional (CNN) management. Python programming tool will be used well GUI that facilitate human–computer interactions. tested result evaluated assess functionality adequacy system.

Language: Английский

---