Anonymity algorithm-based data sharing privacy protection method

An anonymity algorithm-based data sharing privacy protection method, comprising:

• identifying privacy-related attributes of a structured medical data set;
• performing semantic standardization and merging on the identified attributes by a MSAK anonymity algorithm, constructing a multi-semantic sensitive attribute classification tree, and calculating a minimum sensitive attribute difference;
• constructing a virtual initial clustering center, and performing data set clustering and dividing on the identified structured medical data set and the multi-semantic sensitive attribute classification tree to obtain first subdata sets and a sensitive attribute difference matrix;
• performing outlier detection and separation on the first subdata sets to obtain second subdata sets, and a first tuple set to be allocated and comprising outliers;
• setting a parameter k of a k-Anonymity model and a parameter of a -Diversity model, clustering the second subdata sets, aggregating tuples of the second subdata sets into a clustering cluster, and clipping and determining the clustering cluster according to a size, the parameter k and the parameter of the clustering cluster to obtain a first clustering cluster set, a second clustering cluster set and a second tuple set to be allocated;
• determining whether the k-Anonymity model is met according to a number of the tuples in the clustering cluster;
• according to the sensitive attribute difference matrix, calculating an overall difference of sensitive attribute values of the clustering cluster, and determining whether the overall difference of the sensitive attribute values of the clustering cluster meets the -Diversity model;
• and
• re-allocating the tuples in the tuple set to be allocated to the first clustering cluster set and the second clustering cluster set one by one, merging and adjusting remaining clustering clusters in the first clustering cluster set, and performing generalization and suppression to obtain an anonymized data set.