Traffic Mining Applied to Police Activities

This book presents high-quality original contributions on the development of automatic traffic analysis systems that are able to not only anticipate traffic scenarios, but also understand the behavior of road users (vehicles, bikes, trucks, etc.) in order to provide better traffic management, prevent accidents and, potentially, identify criminal behaviors. Topics also include traffic surveillance and vehicle accident analysis using formal concept analysis, convolutional and recurrent neural networks, unsupervised learning and process mining. The content is based on papers presented at the 1st Italian Conference for the Traffic Police (TRAP), which was held in Rome in October 2017. This conference represents a targeted response to the challenges facing the police in connection with managing massive traffic data, finding patterns from historical datasets, and analyzing complex traffic phenomena in order to anticipate potential criminal behaviors. The book will appeal to researchers, practitioners and decision makers interested in traffic monitoring and analysis, traffic modeling and simulation, mobility and social data mining, as well as members of the police.

Gathers the proceedings of the 1st Italian Conference for the Traffic Police, Rome, Italy, October 25-26, 2017 Aimed at researchers as well as practitioners, standard developers and policymakers Presents high-quality peer-reviewed papers covering various aspects of data traffic analysis for police activities

Inhalt
Advancements in Mobility Data Analysis.- Towards a Pervasive and Predictive Trafc Police.- A Process Mining Trafc Behavior.- Efcient and Accurate Trafc Flow Prediction via Fast Dynamic Tensor Completion.- Unsupervised Classication of Routes and Plates from the Trap2017 Dataset.- Vehicle classication based on convolutional networks applied to FMCW radar signals.- Trafc Data: Exploratory Data Analysis with Apache Accumulo.- Exploiting Recurrent Neural Networks for Gate Trafc Prediction.