Techniques for utilizing data mining technology to extract universal topics with multilingual representations from a multilingual database, and to organize existing or new documents in different languages by analyzing their respective topic distributions.