A Knowledge-and-Data-Driven Modeling Approach for Mobile Payment Mode Recognition

Identifying different mobile payment modes based on cell phone signaling data in user movement is crucial for mobile payment scenario planning, payment risk assessment, and analyzing payment behavior patterns across multiple users. This paper proposes a novel knowledge-and-data-driven modeling approach to optimize the use of signaling data for mobile payment mode recognition. The approach integrates both unparsed signaling data and domain-specific URL- Key knowledge extracted from parsed signaling data that contains payment-related information. First, we develop a data-driven model for mobile payment event recognition, leveraging cell phone signaling data. Next, we establish a knowledge-driven model that utilizes domain URL- Key knowledge to identify mobile payment modes during user movement. These two models are then integrated using a composite coupling operator, and real data from a mobile telecommunication company located in Chongqing are used to calibrate and test our approach. Experimental results show that our approach significantly enhances both performance and efficiency.