| 摘要 |
Provided is a user data classification method, comprising: acquiring several user data and generating several samples on the basis of the user data, each of the samples comprising two user data that have the same identification feature information (S11); on the basis of the distinguishing feature information of the two user data and comparison information of the distinguishing feature information, classifying the samples into samples of the same type, samples of different types or samples of a type to be determined (S12); using a predictive classification model to classify again the samples of the type to be determined into samples of the same type or samples of different types, the predictive classification model being obtained by means of training on the basis of the samples of the same type and the samples of different types and comprehensive description feature information of the user data therein (S13); on the basis of the samples of the same type and the samples of different types, classifying the user data (S14). This enables a more accurate association recognition of user data according to the feature information of the user data, better classifying the user data into corresponding user records, and generalizing multiple pieces of comprehensive description feature information of the user, so as to service the user. |
