XML Privacy Preserving Method Based On Dynamic Context

In the era of big data,the contradiction between data value mining and privacy protection is increasingly prominent.Traditional information security emphasizes on the protection of attribute values and often neglects semantics,while the data privacy protection in the era of big data requires effective data usage without leaking sensitive user information,which emphasizes on security solutions that can exploit the data?s value.Previous research on unstructured data security focuses on the improvement of traditional model methods such as attribute encryption,anonymous publishing,and access control.Some research also considers using the static context to process access control rule and designing the top-level data structure to prevent reasoning attacks.However,in the new situation of rapid development of data patterns and data requirements,the existing data privacy protection model is also facing new challenges.(1)Static access control model does not take into account the dynamic context,which makes it ineffective in a flexible network.(2)The management of the data in the existing scheme depends on the node positioning in the hierarchical structure.But in practical application the path location and condition judgment brought by the redundancy of the tree structure cost a great deal of hardware energy.As the complexity of the data deepens and the amount of data processing increases,it may even lead to the collapse of the engine.(3)Any form of access control schemes that neglects the semantic association between data elements and the associated information of the hierarchy itself,may cause excessive protection of data information.(4)Theoretically it is very hard to verify the completeness of reasoning of the top-level structure,while in practice the prior knowledge accumulated during the dynamic process of continuous query results is often ignored,which may overlook possible background reasoning attack in actual operation.Therefore,combining the semantic nature of big data objects and the semi-structured composition of multi-source data,this paper studies the protection scheme of semi-structured data XML based on dynamic context layered design,which can provide with efficient and maximal usage of data,and at the same time meets the secure needs of original data privacy protection.The specific research works include: Firstly,based on user ontology this paper explored the inherent semantic association of data,and combined existing normalization theory to formally describe ontology-based semantic dependencies,and gave the theoretical system of semantic dependence through reasoning and related research.Secondly,according to the essence of semantic association between multi-source data,following the bottom-up integration process,this paper avoided the traditional redundant structure dependent data model,and proposed a generating algorithm that can get the inverted global view to implement semantic-based access control scheme.What?s more,this paper strengthened users? role classification as semantic subject,combining the principle of “subject negative priority” in access control rules,this paper effectively limited the scope of access to original data,and then applied query strategy to process small-scale data,avoiding privacy attribute access and improving query efficiency.Finally,through introducing the concept of dynamic context which may cause privacy leakage,this paper defined prior knowledge base and privacy bipartite graph,analyzed and researched privacy protection model against background knowledge attacks.Specifically,the contributions and innovations of this paper are as follows:(1)A new method of data organization based on semantic dependence is proposed.In big data,same-name with different-meanings or different-names with same meaning is a common form of data aggregation.Introducing semantic association into multi-source data fusion description can not only solve inconsistency,but also describe the richest semantic information with the simplest structure.(2)The design idea of dynamic context layering is proposed.This paper divides dynamic contexts into different levels according to the different stages of the natural process of individual accessing data: "login"-"query"-"continuous query".Through the dynamic management of different levels of semantics,the privacy protections with different goals are achieved.(3)A new access control model for authorization acquisition based on semantic association is proposed.The semantic association-based access control scheme proposed in this paper can pursue the highest data value without paying attention to the organization of the data,avoiding the redundant structure,and efficiently obtaining the data that the user can authorize.At the same time,the access control rule “negative priority” is combined with the semantic association of the user subject to improve the efficiency of data query under secure access.(4)A new privacy protection model against background knowledge attacks in actual operation is proposed.In this paper,user prior knowledge is introduced into the dynamic context as semantics,which effectively leads to changes in the semantic environment based on the user's prior knowledge accumulation.At the practical level,the scheme for preventing privacy leakage is given.