However, tag is different from the general directory structure classification method. First of all, tag can refine classification with less cost. Imagine that an article covering a wide range, such as an article about the achievements of physics since the 20th century, may involve scientists such as relativity, quantum mechanics, black hole theory, Big Bang theory, Einstein, Planck, and even the Nobel Prize. If the classification method of directory structure is used, it is impossible to classify according to all aspects involved in this paper, because refining the classification will make the whole directory structure extremely huge, which is not conducive to the organization and search of data. On the other hand, labels are different. Regardless of the directory structure, they are free to classify articles. The relationship between tags is parallel, but according to correlation analysis, tags that often appear together can be correlated to produce a correlation classification. 2. Tags can also be said to be keyword tags, which are beneficial to search and search. But tag is also different from ordinary keywords. When searching with keywords, you can only find the keywords mentioned in the article, while tag can mark the keywords that are not in the article at all. For example, I can label the above article as "data" or "history", of course, it is more often labeled as "physics". However, if I label it with "data", I can relate all the informative articles.