Geospatial Modeling of Road Traffic Using a Semi-Supervised Regression Algorithm

Abstract

Nowadays, big cities are facing many challenges with respect to traffic congestion, climate change, air and water pollution, among others. Thus, smart cities are intended to improve the life quality of the citizens, tackling such issues with the integration of information and communication technologies to reduce the impact and achieve a well-being state of citizens. In this work, a model to predict the traffic congestion applying a support vector machine method is proposed. In addition, a crowdsourcing approach based on mining the Twitter social networks collecting events associated with the traffic is also proposed. The main contribution of this research is focused on providing a methodology that characterizes the traffic congestion analyzing crowd-sensed data from a geospatial perspective. This approach was implemented over the Mexico City as a case study, in order to forecast possible future traffic events in the city, in which the citizens share their particular situation to discover alternatives routes for avoiding the traffic congestion. Future works are oriented towards designing mobile applications in order to introduce the proposed approach and integrate information from multiple platforms and navigation systems. INDEX TERMS Geospatial modeling, machine learning, regression, road traffic, urban computing, volunteered geographic information.

Traffic Congestion Analysis Based on a Web-GIS and Data Mining of Traffic Events from Twitter

“…According to [ 10 ], traffic congestion is a problem that directly affects big cities. Different approaches to describe this state have been deeply analyzed.…”

Optimized CapsNet for Traffic Jam Speed Prediction Using Mobile Sensor Data under Urban Swarming Transportation

“…Many SVM-based models have been widely employed for traffic prediction. For example, Zeng et al proposed AOSVR to deal with time efficiency of the traffic flow prediction [17], Saldana-Perez et al took advantage of social media data to characterize the traffic congestion and analyze crowd-sensed data from a geospatial perspective [18]. Yan, H. and D.-J.…”