Evaluation of a data-driven intelligent waste classification system for scientific management of garbage recycling in a Chinese community.

Abstract

Waste classification management is effective in addressing the increasing waste output and continuous deterioration of environmental conditions. The waste classification behaviour of resident is an important basis for managers to collect and allocate resources. Traditional analysis methods, such as questionnaire, have limitations considering the complexity of individual behaviour. An intelligent waste classification system (IWCS) was applied and studied in a community for 1 year. Time-based data analysis framework was constructed to describe the residents' waste sorting behaviour and evaluate the IWCS. The results showed that residents preferred to use face recognition than other modes of identification. The ratio of waste delivery frequency was 18.34% in the morning and 81.66% in the evening, respectively. The optimal time windows of disposing wastes were from 6:55 to 9:05 in the morning and from 18:05 to 20:55 in the evening which can avoid crowding. The percentage of accuracy of waste disposal increased gradually in a year. The amount of waste disposal was largest on every Sunday. The average accuracy was more than 94% based on monthly data, but the number of participating residents decreased gradually. Therefore, the study demonstrates that IWCS is a potential platform for increasing the accuracy and efficiency of waste disposal and can promote regulations implementation.