Multi-dimensional data publishing method and system of local differential privacy based on incremental learning

The invention belongs to the field of data security and privacy protection, and provides a local differential privacy multi-dimensional data publishing method and system based on incremental learning, and the method comprises the steps: learning the correlation of all attribute pairs by aggregating a first batch of user disturbance data; a dependency graph model is constructed according to the correlation of the attribute pairs, and the constructed dependency graph model is converted into a connection tree model composed of a plurality of clusters through a connection tree algorithm; on the basis of the second batch of user data, according to the number and size type of attributes contained in each group, estimating the distribution of the groups by adopting a corresponding estimation method to obtain the joint distribution of each group in the junction tree model; and according to the joint tree model and the joint distribution of each group in the joint tree model, generating a data set which also contains