Abstract:

For the use rate of international roaming communications products are too low,based on full volume data of mobile user in communication behavior and attribute,a method of combining statistical analysis and building recognition model was proposed.Using SPSS and Clementine tools,with high hit rate and wide coverage as the target,the user potential features of the international roaming communications products were extracted.Subsequently,the empirical analysis of the potential features also was given,and the very high accuracy rate was obtained.This has important reference value for operators to carry out international roaming service better.

Key words: international roaming, user behavior characteristics, identification model, empirical analysis