7. Spatio-temporal Data Analytics for e-Waste Management System Using Hybrid Deep Belief Networks

In the most recent few decades, there has been a significant increase all over the world in the amount of waste electronic equipment. This is a result of a number of factors, including an increase in production, management rules that are ineffective, recycling practices that are inefficient, and safety risks that are unacceptable. Because of the harmful emissions that are released into the air, water, and soil when electronic waste is disposed of in landfills, this practice has the potential to be detrimental to both human health and the environment. The process of recycling used electronic equipment results in the production of several distinct categories of secondary waste, including solids, liquids, and gases. When waste, such as electronic waste, is thrown away in an improper manner, it has a detrimental effect not only on human health but also on the environment. In this work, we propose a hybrid deep learning model (HDLM) which comprises Fuzzy-based Spatio-Temporal Optimization Mechanism (FSTOM) in addition to Deep Belief networks (DBN) for e-waste prediction. During network training and error checking, residual learning is designed to prevent oscillations. In this chapter, we propose using a novel algorithm called MSOK (modified self-organizing map and K-Means algorithm) to create a profile for each type of waste produced. To take advantage of the best features of both statistical modelling and deep belief network modelling, these hybrid models combine the two. The findings suggest that cutting-edge data analysis techniques could be employed to obtain more accurate statistics regarding garbage generation. Compared with the existing models, the proposed model performs well effectively in determination and classification of the e-wastes with proper optimization strategies.