An Interval-Valued Neural Network Approach for Uncertainty Quantification in Short-Term Wind Speed Prediction
Résumé
We consider the task of performing prediction with neural networks on the basis of uncertain input data expressed in the form of intervals. We aim at quantifying the uncertainty in the prediction arising from both the input data and the prediction model. A multi-layer perceptron neural network (NN) is trained to map interval-valued input data into interval outputs, representing the prediction intervals (PIs) of the real target values. The NN training is performed by non-dominated sorting genetic algorithm-II (NSGA-II), so that the PIs are optimized both in terms of accuracy (coverage probability) and dimension (width). Demonstration of the proposed method is given on two case studies: (i) a synthetic case study, in which the data have been generated with a 5-min time frequency from an Auto-Regressive Moving Average (ARMA) model with either Gaussian or Chi-squared innovation distribution; (ii) a real case study, in which experimental data consist in wind speed measurements with a time-step of 1-hour. Comparisons are given with a crisp (single-valued) approach. The results show that the crisp approach is less reliable than the interval-valued input approach in terms of capturing the variability in input.
Origine | Fichiers produits par l'(les) auteur(s) |
---|
Loading...