Research On Privacy Protection Algorithm Based On Sensitive-k Anonymity

Due to the rapid development of Internet technology,the channels for mass data to be transmitted between users have increased significantly,providing users with great convenience.Although data providers sharing personal information can bring some convenience to themselves,some criminals can use the shared information to attack users.Therefore,it is necessary to perform privacy protection processing before publishing data tables containing sensitive information,so that attackers cannot deduce the user's private information from the published data,and at the same time,the availability of the data can be guaranteed.Most data has the characteristics of dynamic changes.For example,data with higher or lower sensitivity appears for the first time.For example,the frequency of sensitive attributes exceeds the publishing requirements.If the data is processed in the same way,the attacker will use the data table before and after the change.It is very easy to derive special data.In view of the data with many changing characteristics,this article proposes a set of dynamic data processing strategies.If the data appears for the first time,it will be hidden or inserted according to its sensitivity;if the data does not appear for the first time,it will be inserted or stored according to the frequency of occurrence.In the temporary table;when looking for the equivalence class most similar to the new data,the KL divergence is used to calculate the similarity.And on the basis of this group strategy,a dynamic data-oriented(p,?)-Sensitive k anonymous algorithm is proposed.The analysis of experimental results shows that this algorithm can provide privacy protection for dynamic data while increasing data availability.Traditional k-means clustering only divides the data based on k randomly selected sample points,without considering the characteristics of multiple sensitive attributes,and most privacy protection algorithms are mainly focused on single sensitive attributes.Aiming at the characteristics of data with multiple sensitive attributes,a Sensitive-k anonymous model based on clustering of multiple sensitive attributes is proposed.The model first proposes a k-means clustering algorithm based on multiple sensitive attributes,which converts multi-dimensional sensitive attribute values into corresponding multi-dimensional vectors,and then clusters the original data set into k groups according to the similarity calculation of the vector distance.Finally,based on the clustering of multiple sensitive attributes,a data table that meets the requirements of anonymity is constructed.The experimental results show that the Sensitive-k anonymous model based on multi-sensitive attribute clustering has a good performance in the privacy protection of multiple sensitive attributes.