Generating String Recommendation Efficiently and Privately

Recommender systems have turned into a critical tool for personalization of online services. Producing proposals in online administrations relies on upon privacy-sensitive data gathered from the clients. Conventional information assurance mechanisms concentrate on access control and secure transmission, which give security just against malicious third parties, yet not the service provider. This makes a genuine security hazard for the clients. In this paper, we expect to ensure the private information against the service provider while saving the usefulness of the framework. We propose encoding private information and preparing them under encryption to create suggestions. By presenting a semi-trusted third party and utilizing data packing, we develop an exceedingly proficient framework that does not require the dynamic interest of the client. We additionally exhibit a comparison protocol, which is the first to the best of our insight that analyzes different qualities that are packed in one encryption. Directed trials demonstrate that this work opens a way to generate private recommendations in a privacy-preserving manner. The existing system work on only the integer recommendation but in our propose work we implement on phrase and string recommendation by applying steaming and stop word removal.