Multi-data source data classification method and server

The invention discloses a multi-data source data classification method, which is used for solving the problems of non-unified classification tag standards and information classification disorder of anexisting multimedia information platform. The method provided by the invention comprises the steps of determining target data in data displayed by the multimedia information platform according to a preset mode; extracting keywords from the target data; performing statistics on first word frequencies of the keywords occurring in title positions of the target data and second word frequencies occurring in content positions; according to the first word frequencies, the second word frequencies, corresponding first weights and corresponding second weights, calculating word popularity values of thekeywords; determining the keyword with the highest word popularity value as a new classification tag; and performing matching on the target data by adopting the new classification tag, and applying the new classification tag t