Efficient Multi-keyword Approximate Search on Encrypted Cloud Data

• Main Authors: WAN-NI SHIH, 施宛妮
• Other Authors: Tai-Lin Chin
• Format: Others
• Language: en_US
• Published: 2017
• Online Access: http://ndltd.ncl.edu.tw/handle/yezd9k

碩士 === 國立臺灣科技大學 === 資訊工程系 === 105 === Now with the growing popularity of cloud computing, data owners outsource the data to cloud for convenience management. Despite the advantages of cloud storage, putting a lot of attention on data security is definitely necessary. Before outsourcing the sensitive data, data owners should encrypt it for privacy requirements. However, the encryption could make search problem of utilization become difficult by keywords. There are some existing techniques provide keyword search over encrypted data. Those strategies contain weakness about the inefficient searchable vectors. The previous strategies maintain a large number of words as the elements to construct the index for each document. It will cause the amount of calculation that sharply increase. In this thesis, we propose a novel keyword search over encrypted data scheme by reducing the size of the index vectors. First, we analyze data through latent semantic analysis, to discover the association between different words. Then, we select the principal dimension to retain the important information of dataset. After transforming the data from high-dimensional space to low-dimensional space, we provide appropriate result to the user's request. Our proposed scheme support approximate keyword search with premeditate data analysis rather than maintain large number of words. Experiments on real-world dataset show that our scheme not only provides flexibly query for users, but also effectively reduce the amount of calculation.