The sensitivity of the traffic detector or the influence of the external environment will cause different degrees of data missing or abnormal. In addition, the traffic parameters collected by different traffic detectors are different. Therefore, it is very important to identify key data sources for multi-source data fusion for different research purposes. This paper establishes a key data source evaluation system, and puts four evaluation indicators of data integrity, data dispersion, data repair ability and data support degree, and calculates the index weight by the entropy weight method. Taking the geomagnetic data, video bayonet data and floating car data of Jinan City as examples, taking the traffic situation discrimination as the research purpose, selecting the travel speed as the fusion index, the case analysis is carried out to verify the effectiveness of the key data source identification method proposed in this paper.

CICTP