This paper introduces two new closely related betweenness centrality measures based on the Randomized Shortest Paths (RSP) framework, which fill a gap between traditional network centrality measures based on shortest paths and more recent methods considering random walks or current flows. The framework defines Boltzmann probability distributions over paths of the network which focus on the shortest paths, but also take into account longer paths depending on an inverse temperature parameter. RSP’s have previously proven to be useful in defining distance measures on networks. In this work we study their utility in quantifying the importance of the nodes of a network. The proposed RSP betweenness centralities combine, in an optimal way, the ideas of using the shortest and purely random paths for analysing the roles of network nodes, avoiding issues involving these two paradigms. We present the derivations of these measures and how they can be computed in an efficient way. In addition, we show with real world examples the potential of the RSP betweenness centralities in identifying interesting nodes of a network that more traditional methods might fail to notice.

No abstract

No abstract

Road traffic forecasting is crucial in Intelligent Transportation Systems (ITS). To achieve accurate results, it is necessary to model the dynamic nature and the complex non-linear dependencies governing traffic. The goal is particularly challenging when the prediction involves more than just one traffic variable. This paper proposes a novel multi-task learning model, called AST-MTL, to perform multi-horizon predictions of the traffic flow and speed at the road network scale. The strategy combines a multilayer fully-connected neural network (FNN) and a multi-head attention mechanism to learn related tasks while improving generalization performance. The model also includes the graph convolutional network (GCNs) and the gated recurrent unit network (GRUs) to extract the spatial and temporal features of traffic conditions. Our experiments employ new sets of GPS data, called OBU data, to perform traffic prediction in the freeway and urban contexts. The experimental results prove our model can effectively perform multi-horizon traffic forecasting for different types of roads and outperform state-of-the-art models.INDEX TERMS Deep learning, multi-task learning, graph mining, traffic prediction.

No abstract