A secure neural network‐based ranking approach for document searching in cloud data center. (17th May 2022)

• Record Type: Journal Article
• Authors: Malhotra, Sheenam; Singh, Williamjeet
• Is Part Of: Software, practice & experience. Volume 52:Number 9(2022)
• Page Start: 1927
• Page End: 1946
• Publication Date: 2022-05-17
• Subjects: artificial neural network -- cloud computing -- inverse document frequency -- ranking -- symmetric searchable encryption; Computer software -- Periodicals; Computer programming -- Periodicals; Computer programs -- Periodicals; 005.3
• DOI: 10.1002/spe.3097
• Languages: English
• ISSNs: 0038-0644

Abstract

Abstract: Cloud computing has gained attention due to its sophisticated processing architecture and data storage capabilities in the last couple of years. Due to high volume data and the endless number of possible users, the security of the account holder's stored data and privacy becomes essential for this paradigm. This article focuses on the encryption architecture of data storage and retrieval by creating an encrypted searchable index, which is inspired by symmetric searchable encryption. Ranking becomes a need for providing the best out of the search results to the user. This research article proposes an efficient and flexible artificial neural network (ANN) based ranking scheme to search for documents from the cloud server. The proposed algorithm architecture is segmented into three parts. The first part is the generation of the encryption index over the uploaded data, the second part is query analysis, and the third part is ranking. To consolidate the encryption mechanism, RSA, NTRU, and AES were used based on the requirement of the data. To orient the retrieval part, the degree of the top keyword in the server is determined by using term frequency with inverse document frequency schemes. The retrieved documents are further ranked using ANNs. The simulation setup was done on MATLAB 2016b having datasets from Kaggle (Twitter data) and FIRE dataset.