Evaluation of artificial neural networks and the ARIMA model applied to temperature prediction in a charcoal oven

Descrição

Tipo

Artigo de periódico

Título alternativo

Primeiro orientador

Membros da banca

Resumo

This work used the Artificial Neural Networks RBF- Radial Basis Function and MLP- Multilayer Perceptron, the Neurofuzzy ANFIS- Adaptive-network-based Fuzzy Inference System and the model ARIMA-Auto Regressive Integrated Moving Average, to predict the temperature. Different topologies were tested for Artificial Neural Networks and also for Neurofuzzy, in order to find the best configuration. To evaluate the performance of the predictors, the following metrics were used: Root Mean Square Error - RMSE, Mean Square Error - MSE and Mean Absolute Error - MAE. The analysis of the quality of the predictors was also performed using Theil's U coefficient. Analyzing the results of this study, it was possible to verify that the Artificial Neural Networks, the Neurofuzzy Network and the ARIMA model, can be applied to predict the temperature in charcoal kilns, with the MLP and ANFIS networks presenting the best results.

Abstract

Assunto

Redes neurais (Computação), Carvão vegetal, Controle de temperatura

Palavras-chave

Automatic temperature forecast, Methods, Performance

Citação

Curso

Endereço externo

http://www.ijern.com/journal/2023/March-2023/04.pdf

Avaliação

Revisão

Suplementado Por

Referenciado Por