Publicação
A tree-based approach to forecast the total nitrogen in wastewater treatment plants
| Resumo: | With the increase in the world population, there has been an increase in environmental problems worldwide. One of these problems is the quality of the water, which can cause problems to society’s well-being and the environments surrounding it. Wastewater Treatment Plants (WWTPs) emerged to address this problem. It is necessary to pay attention to the different substances present in the waterwaters treated in the WWTPs, as is the case of total nitrogen, which can cause severe damage to the environment. Therefore, this work aims to forecast the total nitrogen in a WWTP by conceiving, tuning and evaluating several Machine Learning (ML) models, particularly the Decision Trees (DTs) and the Random Forest (RF). The best candidate model was a DT-based with an approximate error of 1.6 mg/L. Considering the best candidate model identified, our objective was to extract the rules generated by the model to understand the factors that lead to high values at the level of total nitrogen. |
|---|---|
| Autores principais: | Faria, Carlos |
| Outros Autores: | Oliveira, Pedro; Fernandes, B.; Aguiar, Francisco; Pereira, M. A.; Novais, Paulo |
| Assunto: | Engenharia e Tecnologia::Engenharia Eletrotécnica, Eletrónica e Informática |
| Ano: | 2022 |
| País: | Portugal |
| Tipo de documento: | comunicação em conferência |
| Tipo de acesso: | acesso restrito |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | With the increase in the world population, there has been an increase in environmental problems worldwide. One of these problems is the quality of the water, which can cause problems to society’s well-being and the environments surrounding it. Wastewater Treatment Plants (WWTPs) emerged to address this problem. It is necessary to pay attention to the different substances present in the waterwaters treated in the WWTPs, as is the case of total nitrogen, which can cause severe damage to the environment. Therefore, this work aims to forecast the total nitrogen in a WWTP by conceiving, tuning and evaluating several Machine Learning (ML) models, particularly the Decision Trees (DTs) and the Random Forest (RF). The best candidate model was a DT-based with an approximate error of 1.6 mg/L. Considering the best candidate model identified, our objective was to extract the rules generated by the model to understand the factors that lead to high values at the level of total nitrogen. |
|---|