Propuesta metodológica para la optimización de modelos predictivos de generación de residuos sólidos municipales en zonas urbanas
Date
2024-06-19
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Pontificia Universidad Católica del Perú
Abstract
El pronóstico de la generación de residuos sólidos municipales (RSM) desempeña un papel
esencial en la toma de decisiones y proporciona información relevante para la gestión de
residuos, así como una comprensión profunda de los factores que influyen en este proceso. En
este trabajo, se desarrolló un modelo de predicción de RSM específico para Lima
Metropolitana, basado en variables socioculturales, ambientales y económicas, teniendo al
2019 como año de referencia, debido a la influencia del COVID-19 en los datos sobre este tema
en años posteriores a la pandemia. El modelo se construyó utilizando las cantidades per cápita
de RSM generadas en cada distrito, junto con parámetros relacionados con el consumo de
combustibles en el hogar (como gas natural, electricidad y gas licuado de petróleo) y
características demográficas de la población (como edad, nivel de educación y gasto mensual).
Dada la calidad y disponibilidad de datos, se optó por utilizar el algoritmo de random forest
como técnica de predicción. Las variables analizadas se obtuvieron a partir de la Encuesta
Residencial de Consumo y Uso de Energía (ERCUE) a nivel municipal. Los resultados
indicaron que el algoritmo implementado explica el 51% de la variabilidad de los datos. Se
espera que las recomendaciones presentadas en este estudio sirvan para investigaciones futuras
relacionadas con la predicción de RSM, contribuyendo a obtener resultados más precisos y
aplicables a contextos específicos.
Municipal solid waste (MSW) generation forecasting plays an essential role in decision making and provides relevant information for waste management, as well as a deep understanding of the factors that influence this process. In this work, a specific MSW prediction model was developed for Metropolitan Lima, based on sociocultural, environmental and economic variables, having 2019 as the reference year, due to the influence of COVID-19 on data on this topic in post-pandemic years. The model was constructed using per capita amounts of MSW generated in each district, along with parameters related to household fuel consumption (such as natural gas, electricity, and liquefied petroleum gas) and demographic characteristics of the population (such as age, education level, and monthly expenditure). Given the quality and availability of data, we chose to use the random forest algorithm as a prediction technique. The variables analyzed were obtained from the Residential Survey of Energy Consumption and Use (ERCUE) at the municipal level. The results indicated that the implemented algorithm explains 51% of the variability of the data. It is expected that the recommendations presented in this study will be useful for future research related to MSW prediction, contributing to obtain more accurate results applicable to specific contexts.
Municipal solid waste (MSW) generation forecasting plays an essential role in decision making and provides relevant information for waste management, as well as a deep understanding of the factors that influence this process. In this work, a specific MSW prediction model was developed for Metropolitan Lima, based on sociocultural, environmental and economic variables, having 2019 as the reference year, due to the influence of COVID-19 on data on this topic in post-pandemic years. The model was constructed using per capita amounts of MSW generated in each district, along with parameters related to household fuel consumption (such as natural gas, electricity, and liquefied petroleum gas) and demographic characteristics of the population (such as age, education level, and monthly expenditure). Given the quality and availability of data, we chose to use the random forest algorithm as a prediction technique. The variables analyzed were obtained from the Residential Survey of Energy Consumption and Use (ERCUE) at the municipal level. The results indicated that the implemented algorithm explains 51% of the variability of the data. It is expected that the recommendations presented in this study will be useful for future research related to MSW prediction, contributing to obtain more accurate results applicable to specific contexts.
Description
Keywords
Aprendizaje automático (Inteligencia artificial), Combustibles--Consumo, Residuos sólidos
Citation
Collections
Endorsement
Review
Supplemented By
Referenced By
Creative Commons license
Except where otherwised noted, this item's license is described as info:eu-repo/semantics/openAccess