The extraction of customer´s behavioral association rules under the variable granularity based on granular computing

This paper tries to explore the information needs at different management levels from the perspective of granular computing. Firstly, it introduces the concept of information granularity, granularity measurement, granularity structure, conversion method between different granularity levels, and the extraction of association rules under the variable granularity. And then it improves the existing single attribute granularity definition, and puts forward a new measurement method of information granularities with multiple attributes. Based on the conversion method between different granularity levels, this paper lists some proposals for the extraction of association rules on customer behavior under variable granularity. Finally, through empirical analysis, this paper has discussed how to simulate the conversion process between different granularity levels, and get the information granule as well as regular features.