FRIEND RECOMMENDATION FOR EFFICIENT NOVEL SECURE COMMUNICATION USING ONLINE SOCIAL NETWORK

Abstract

Recommendation system (RS) plays a vital role for the online data users. But most of the approaches based on the RS in online public networks are not reliable with the interest of data users. In order to overcome the untrustworthiness and uncertainty of user based RS in public network. We have proposed a Novel Secure Key Distribution based Recommendation System (NSKD-RS) for preserving the privacy of the data users using neural networks. Data User (DU) send request to other DU through the Trusted Third Party (TTP). TTP will search the similar interest DUs based on the received search request. The DUs interest is placed in terms of tag. If the DU interest matches with other DU interest then two DU’s will become friends by using neural networks. The cloud has the collection of the data user’s files, here the neural networks are used for the data classification based on the data user’s interest in addition to neural networks provide the recommended users based on the similar interest. We have proposed the secure recommendation system based key distribution mechanism for the tag decryption as well as to have secure communication between the DUs. The secure tag matching process is developed in order to reduce the computation complexity as well as to have the secure data sharing between the DUs. By using our proposed NSKD-RS the commutation as well as the communication complexity is reduced.