Article Preview
TopLiterature Survey
The author talks about different networks and finds the relationship to recommend friends. It has two major components first being related networks by selecting important features and the second being network structure and preserving most of it. It is based on friend correlation and considers effect in different social roles. Huang et al. (2016).
The author gives idea about a new friend recommendation system using artificial bee colony(ABC) which indicates a link between users. It is based on the structural properties of the social network. Firstly, it finds the relevant parameters for the relationship among users using social topology. The sub- graph of the network is composed of users and all users within the network separated by three degrees of separation, then based on the subgraph new links are suggested thus indicating new friends. Akbari et al. (2013).
The paper gives a idea about a more precise friend recommendation with 2 stages. In the first stage the information of the relationship between text and users, then align the recommended friends. In the second stage, they built the topic model of the relationship between image features and users. Huang et al. (2017).
The paper proposed a novel semantic-based recommendation of friends based on their lifestyles. They take advantage of smartphones, friend books to discover the lifestyle of a user from user-centric data and then measure the similarity of lifestyle to finally recommend from with similar lifestyles. The friend book finally keeps a list of people with the highest score to recommend a friend. Wang et al. (2015).
The paper gives idea about users who want to meet friends on social media, they interviewed active users and then developed a friend request acceptance model to refer to various factors that influence it. They found out the major factor that impacts the person who accepts the friend request, mostly person with common hobbies and mutual friend is accepted. Rashtian et al. (2014).
The paper talks about how to find short paths between users in a network which would indicate their closeness and will also result in them being a good recommendation as a friend. These are based on email contact, it is found that the kind of things people talk or find is a huge factor in determining their closeness. Adamic and Adar (2005).