云存储使用户可以将其敏感数据外包存储到云服务器上,从而方便数据管理.然而数据存储在第三方服务器后,用户对其不再完全可控,从而对其数据隐私的保护则显得极其重要.所以,用户通常会先对其数据进行加密处理,再存储到云服务器上,从而保证数据的私密性.但是数据加密后,用户需要查询相关文件时,对密文数据的检索就是一个迫切需要解决的问题.因此,本文提出一个密文云存储环境下的基于相似搜索平衡二叉树的支持排序的多关键词检索方案 (similar multi-keyword ranked searchable scheme,SMRSS).其核心思想是在构建索引时,对文档集中的每一个文档进行预处理,将每个文档相似度最高的Top-k个文档进行汇集处理,构造出对应的相似搜索平衡二叉树,从而在进行搜索时可以以最快的速度获取Top-k个相似度最高的文档.同时,本文方案利用向量空间模型、TF-IDF加权统计方法来构建索引,并利用欧式距离来确定相似文档以提高对平衡二叉树的搜索效率.由于本文方案的索引结构是基于平衡二叉树,所以可以实现亚线性的搜索时间复杂度.从理论分析及实验结果可以得知,与EDMRS方案相比,SMRSS方案在保证检索准确度和安全性的前提下,极大地提高了检索效率.%Cloud storage enables users to outsource their sensitive data to the cloud server to facilitate data management. However, when the data is stored in the third party server, it is no longer fully controllable. So the protection of data privacy is extremely important. Usually, users encrypt their data first and then store it on the cloud server to ensure data privacy. However, after data is encrypted, the traditional plaintext keyword search is unavailable. In this paper, we propose a similar search balanced binary tree based multi-keyword ranked searchable encryption scheme (SMRSS). The core idea is to construct similar file sets and stores the similar Top-k files with high relevant score in the leaf node of the index tree. When users query for Top-k files, it can get the highest similar files at the fastest speed. Meanwhile, it utilizes vector space model and TF-IDF weighted statistical method to construct an index, and uses Euclidean distance to determine similar files. As the index structure is based on the balanced binary tree, the search complexity is sub-linear. The experimental results show that, compared with the EDMRS scheme, the SMRSS scheme greatly improves the search efficiency while ensuring equivalent accuracy and security.
展开▼
