GRNN Model With Feedback Mechanism Incorporating k-Nearest Neighbor and Modified Gray Wolf Optimization Algorithm in Intelligent Transportation

Xunjin Wu, Jianming Zhan, Weiping Ding, Witold Pedrycz

Abstract: With the rapid expansion of information technology, the complexity of transportation systems is increasing, which promotes the application of prediction on the field of intelligent transportation. However, the existing methods about intelligent transportation prediction have the phenomenon of losing some important information in the prediction process. Meanwhile, some redundant features affect the improvement of prediction accuracy. Moreover, the randomness of k-value selection in k-nearest neighbour (KNN) leads to some limitations in the prediction performance of intelligent transportation prediction. To address these challenges, this paper designs a dynamic prediction system of generalized regression neural network (GRNN) with feedback mechanism by virtue of KNN and modified grey wolf optimization algorithm (MGWO), named KNN-MGWO-FMGRNN. Firstly, the discrepancy value between different samples for each feature is calculated and the k nearest samples to a certain sample are taken as the discrepancy result of this sample under this feature. Then, the KNN results are fused with the difference results and combined with the MGWO algorithm to obtain the optimal k-value. Next, the optimal k-value is used to recalculate the difference result and the KNN result to complete the feature subset selection (FSS). Further, the optimal FSS and model learning are completed simultaneously using FMGRNN. Finally, the availability and merits of the established model are affirmed by six real data sets.