PROJECT TITLE :

Improving the Quality of Prediction Intervals Through Optimal Aggregation

ABSTRACT:

Neural Networks (NNs) are a good tool to model nonlinear systems. However, their forecasting performance considerably drops in the presence of method uncertainties and disturbances. NN-based mostly prediction intervals (PIs) supply an alternate resolution to appropriately quantify uncertainties and disturbances associated with point forecasts. In this paper, an NN ensemble procedure is proposed to construct quality PIs. A recently developed lower–upper certain estimation method is applied to develop NN-primarily based PIs. Then, made PIs from the NN ensemble members are combined using a weighted averaging mechanism. Simulated annealing and a genetic algorithm are used to optimally alter the weights for the aggregation mechanism. The proposed method is examined for three different case studies. Simulation results reveal that the proposed method improves the typical PI quality of individual NNs by twenty twopercent, eighteen%, and 78percent for the first, second, and third case studies, respectively. The simulation study additionally demonstrates that a threepercent–4p.c improvement in the quality of PIs will be achieved using the proposed methodology compared to the straightforward averaging aggregation method.