Machine Learning-Based Anomaly Detection for Efficient Waste Management and Disposal: A Comparative Study of Logistic Regression and Random Forest

Abstract

Waste management and disposal have turned out to be essential environmental challenges because of speedy urbanization and increased waste technology. Inefficient waste management, unlawful dumping, and mismanagement can cause intense environmental pollution and health risks. Traditional monitoring systems are regularly guided, labor-intensive, and incapable of coping with large-scale data effectively.This look at proposes a device getting to know-primarily based framework for detecting anomalies in waste control systems. A simulated dataset containing waste kind, amount, series frequency, and disposal patterns is used. Logistic Regression and Random Forest fashions are carried out and evaluated in terms of precision, remember, F1-score, confusion matrix, and ROC curve. The consequences exhibit that Random Forest significantly outperforms Logistic Regression, reaching higher accuracy, and do not forget. They have a look at highlighting the effectiveness of machine learning in improving waste monitoring systems and assisting sustainable environmental management supervised by mastering algorithms—Logistic Regression and Random Forest are carried out and evaluated using metrics which include precision, recall, confusion matrix, and ROC curve.The results imply that the Random Forest version outperforms Logistic Regression in detecting irregularities in waste disposal styles. The findings highlight the potential of system getting to know techniques in improving waste monitoring structures, reducing environmental risks, and assisting sustainable waste management practices.