Top-K Similarity Search over Encrypted Medical Data Set for Disease Evaluation

Peter Morris, Ghilby Jaison

Abstract: The Data owners using big Data applications outsource their data on cloud servers for the convenience of data management and mining. To protect privacy of sensitive data, documents are encrypted and uploaded to the cloud servers. Searchable encryption is used to encrypt information in documents so that the information can be queried by the authorized users. Symmetric encryption is used here. Based on the relevance of words in search query with the term frequency and inverse document frequency, the documents are retrieved from cloud and provided to the authorized users. The proposed system uses Winner Takes All (WTA) concept of LAMSTAR (Large Memory Storage And Retrieval) to classify and retrieve the documents as per the query requested by the authorized users. Here LAMSTAR is used for medical diagnosis problems that concern large data sets of many categories.

Keywords: cloud computing, privacy preserving, multi-keywords search, disease evaluation, similarity search, Lamstar algorithm