Abstract:

Now most of the data published in digital library is of  book resources, however, data with user attributes and library resources is rarely published. The analysts therefore cannot obtain much information from existing released data, which causes problems for some scientific research. We propose an anonymous method which publishes user attributes and book information together. We first recode all books using Chinese library classification numbers, divide the whole data according to the sparse conditions of recoding,  and use the substitution method to anonymously encode in each partition. Experimental results show that the final data has higher accuracy and practicability, and that it is able to directly identify the relationship between attributes via the scatter plot, thus providing more useful information for scientific research.

Key words: digital library, data published, privacy protection, (k,e)anonymity, scatter diagram