Bus IC Card Swiping Behavior Recognition Based on Multivariate Data Fusion and Venn Diagram

摘要

In view of the problem that bus IC card data cannot be identified and separated due to cross-section card swiping, substituted swiping and missing swiping, this paper proposes a special algorithm of bus IC card swiping behavior recognition based on multivariate data fusion and Venn diagram. It is based on the existing data of swiping card of a single passenger on the bus at the same time, data of regular bus operation information, data of regular bus stop and data of boarding stations for passengers who swipe their card. Firstly, the data is preprocessed to delete the noise data in the source data which is caused by GPS missing due to weather and the card reader fault that causes the same card to be swiped many times. And data matching is based on the card swiping data of the same card number, transaction date and the same train number within the specified time threshold are matched with the data of boarding, and the time difference threshold of transaction time is set as 30 seconds. Secondly, through this process, the cross- section card swiping, generation card swiping and missed card swiping population are identified and separated, and the number of population and IC card number are counted. Using the Venn diagram, we can clearly demonstrate the structural relationship between cross-section card swiping, substituted swiping and missing swiping. Finally, the data of passengers on and off the bus are completed to obtain correct passenger flow analysis and prediction data, and the revenue analysis is carried out. After identifying 8,163,017 IC card swiping data of cross-section buses in Xiamen city in November 2018 according to the method in this paper, a total of 5,123,694 people used IC card to take cross-section buses, and 1,503,237 people were found to have swiped card across sections, 506,284 people were swiped on behalf of others, and 289,550 people missed data.