Identifying Toxic Components in E-Waste using IoT and Naive Bayes Classification

Electronic waste (e-waste) comprises several toxic substances, including lead, mercury, and cadmium, which present significant environmental and health hazards if not adequately recognized and disposed of. This paper presents an intelligent system that combines Internet of Things (IoT) sensors with a Naive Bayes classification algorithm to identify hazardous substances in e-waste for secure recycling. Data was gathered via IoT-enabled sensors that measured component kind, material composition, and embedded IDs. The system was trained and verified using a publicly accessible dataset from the e-waste project repository on GitHub, including 3,000 samples of e-waste objects with annotated hazardous and non-hazardous characteristics. The Naive Bayes classifier achieved an overall accuracy of 98.67%, with a precision of 97.28%, a recall of 97.32%, and an F1-score of 97.22%. The model exhibited consistent real-time classification efficacy, with an average decision latency of 1.8 seconds. The proposed IoT-ML system facilitates the prompt identification and classification of hazardous substances in discarded electronics, aiding in automated sorting and sustainable e-waste management. This method improves regulatory adherence and reduces ecological damage linked to inappropriate e-waste disposal.