Examinando por Autor "Duarte Sandoval, Carolina"
Mostrando 1 - 1 de 1
Resultados por página
Opciones de ordenación
- ÍtemElaboración y comparación del pronóstico de la temperatura mínima mediante un modelo ARIMA y redes neuronales de un cultivo de rosa de la sabana de Bogotá(Fundación Universitaria Los Libertadores. Sede Bogotá., ) Duarte Sandoval, Carolina; Sandoval Rodriguez, WilsonThe floriculture sector is very important in the Bogotá´s Savannah due to the high productivity and economic competition in the country. As one of the most developed agricultural sectors, it requires technologies that help to continuously improve the process. Negative changes in the climate contribute to the affectation in productivity, so it is of great importance to know these possible changes in climatic variables such as precipitation, humidity, and temperature. The aim is to obtain a good model with which to forecast changes in minimum temperatures for the correct physiological development of the rose. The forecast was carried out with a database of a floricultural company in the Bogotá´s Savannah and consisted of obtaining a forecast of minimum temperatures for the month of February 2022. Three forecasts were obtained, one with an ARIMA model, one with feedback neural networks and one with recurrent neural networks, which were compared with the real values. The results indicate lower error in the ARIMA model but greater variation in the minimum temperatures with the recurrent neural networks, so it is considered necessary to develop a multivariate series model in which variables such as maximum temperature and relative humidity are included so that the model is more accurate to the real data. KEYWORDS: Forecast, weather, rose, minimum temperatura.