Transfer Learning based architecture for urban transportation Big data fusion

Abstract

Recently, intelligent transportation system (ITS) is considered as one of the most important issues in smart city applications. Its supports urban and regional development and promotes economic growth, social development, and enhances human well-being. ITS integrates new technologies of information and communication including sensors, social media IoT devices which can generate a massive amount of heterogeneous and multimodal data known as big data term. In this context, Data Fusion techniques (DF) seem promising and have emerged from transportation applications and hold a promising opportunities to deal with imperfect raw data for capturing reliable, valuable and accurate information. Literature. In literature many DF techniques based on machine learning remarkably renovates fusion techniques by offering the strong ability of computing and predicting. In this paper, we propose new Hybrid method based on TL (transfer learning) combine tow pertained DL models such as irregular CNN [1], and bi-directional LSTM [2] models to fuse multimodal and spatial temporal data. the propose method use CBOA algorithm for feature selection in to order to provide effective exploration of significant features with faster convergence. The proposed model demonstrated its effective results on the applied dataset by offering good results and outcome over traditional methods.