Research on Distance based Data Records Matching Method for Mobile Network CDR

In the era of big data, telecom operators usually need to vaguely associate data sets generated by different systems or the same system working on different situation. This paper proposes a distance based method for fuzzy association of data records from two different data sets, where distance refers to most of the dimensions that can be used for measurement. With this method, we calculate the non-error distance introduced by the differences of data generation systems, and eliminate the interference of the non-error distance. Then we perform full association on the data sets to achieve a set of records grouped by keys, each group sorted by distance. Finally, we choose the record with minimum distance from each group. With proper keys and dimension, the method described here can be used in many fields such as 4G/5G mobile network detail records association and so on.