Segmentation of municipal solid waste using artificial neural networks

Abstract

Abstract The article discusses the prospects of using neural networks and Waste-to-energy technology to create a rational and efficient waste management culture. The study determines the quality (by metrics) of a trained neural, network that determines the type of solid household waste, depending on various parameters of the model. Based on the analysis of the obtained metrics, a conclusion is made about the best parameters for the developed neural network model. This neural network was trained specifically for this study, and as was chosen TACO dataset. Brief theories of neural networks and Waste-to-energy technologies are also discussedenergy. Particular attention is paid to the need to use these tools together to reduce and suspend the formation of new landfills and energy generation. The article will be especially relevant for scientists in those countries where the percentage of recycled waste tends to zero.