Research On Answering Why-not Questions With Privacy Protection

In the era of big data,the variety of query semantics and performance of data query services has been greatly improved in recent years,while higher requirements are put forward for the data usability of data query services.For enjoying the highly usable data query services as well as reducing the administrative costs,data owners are motivated to outsource their data services to public clouds,however,it may cause serious privacy issues.Through an in-depth analysis of the privacy-preserving requirements on answering why-not questions in the data outsourcing environment,we define and solve the problem of answering why-not questions on top-k queries with privacy protection.To this end,we first propose a basic method to achieve answering why-not questions on top-k queries with privacy protection.In the basic method,a sample-based secure weighting space generation method is presented for obtaining the best approximate refined query.we proposes an optimized method on answering why-not questions on top-k queries with privacy protection.In this method,exploiting the original dominance between data objects,we propose a dominance-based secure pruning.Besides,based on the features of the penalty computation,two early stopping conditions are presented to improve the performance of traversing the weighting space.In addition,we state the proposed methods on answering why-not questions can be adopted to the secure weighted kNN queries.For the above research works,we make a theoretical analysis on the security and computational complexity of the proposed methods.At the same time,over several real datasets,we evaluate the query performance of the proposed methods in the query response time.