Automated Machine Learning in Waste Classification: A Revolutionary Approach to Efficiency and Accuracy
Zhuoxuan Lee, Ying Cheng Wu, Xukang Wang
Abstract
In recent years, waste segregation, treatment, and recycling have become critical global issues, drawing significant attention worldwide. However, the efficiency of recycling processes can be influenced by numerous factors. Among these, waste classification plays a crucial role, and researchers have explored the application of machine learning to automate this step. Nonetheless, traditional machine learning approaches often require skilled professionals to invest substantial time in debugging, and achieving satisfactory accuracy can be challenging. To address this concern, we propose leveraging Automated Machine Learning (AutoML) to enhance the speed and accuracy of waste classification, thereby expediting the implementation of waste segregation policies. Our findings indicate that AutoML outperforms traditional machine learning models, requiring less time and energy, while achieving an impressive accuracy and precision rate of 95.1%.