Predicción de la radiación solar en Iquitos utilizando modelos de redes neuronales artificiales y de series temporales
This study investigated the predictive capability of Artificial Neural Network (ANN) models and time series statistical models in predicting solar radiation in Iquitos. The problem was formulated to address the need for accurate predictive tools in the context of solar energy, a key renewable source...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2024 |
| País: | Perú |
| Institución: | Universidad Nacional De La Amazonía Peruana |
| Repositorio: | UNAPIquitos-Institucional |
| Idioma: | español |
| OAI Identifier: | oai:repositorio.unapiquitos.edu.pe:20.500.12737/11266 |
| Acceso en línea: | https://hdl.handle.net/20.500.12737/11266 |
| Access Level: | acceso abierto |
| Palabra clave: | X https://purl.org/pe-repo/ocde/ford#2.02.04 |
| Sumario: | This study investigated the predictive capability of Artificial Neural Network (ANN) models and time series statistical models in predicting solar radiation in Iquitos. The problem was formulated to address the need for accurate predictive tools in the context of solar energy, a key renewable source. The objectives were focused on comparing these two approaches to identify the most effective in terms of accuracy and reliability. The methodology involved the collection of historical meteorological data, followed by the implementation and comparison of both models using indicators such as Mean Squared Error (MSE) and the correlation coefficient (R). The results revealed that, in the testing phase, the ANN model achieved an MSE of 4.3404 and an R of 0.95, surpassing the time series model, which registered an MSE of 647.8323 and an R of 0.83 in the same phase. We conclude that Artificial Neural Networks provide a more accurate and reliable approach to predicting solar radiation in Iquitos compared to time series models. This finding underscores the potential of advanced machine learning techniques in climate prediction and their relevance for energy and environmental planning. However, it is recognized that no model is universally superior, and the choice should be based on a combination of accuracy, reliability, interpretability, and adaptability to specific data. |
|---|