A B S T R A C T
Link prediction is a key technique in many applications in social networks; where potential links between entities need to be predicted. Typical link prediction techniques deal with either uniform entities, i.e., company to company, applicant to applicant links, or non-mutual relationships, e.g., company to applicant links. However, there is a challenging problem of link prediction among the composite entities and mutual links; such as accurate prediction of matches on company dataset, jobs or workers on employment websites, where the links are mutually determined by both entities that composite entity belong to disjoint groups. The causes of interactions in these domains makes composite and mutual link prediction significantly different from the typical version of the problem. This work addresses these issues by proposing the Support Vector Machine model. By implementing the proposed algorithm it is expected that the accuracy will get increased in the link prediction problem.
Key words: Link prediction, Potential links, Composite, Mutual links, Support Vector Machine.
INTRODUCTION
A social networking service is an online service, platform, or site that target to facilitating the building of social networks or social relations between people who, for example, share importance's, activities, backgrounds, or real-life communications. A social network service consists of a representation of each user, his/her social links, and a variety of new services. This is used to model the interaction among the communities on the social networks. Where the graphs are used to represent the interactions between those communities, in which nodes are representing to people in some communities and links are representing the association between those people.
Understanding the association between two specific nodes by predicting the likelihood of a future but not currently existing association between them is a fundamental problem known as link prediction.
Interaction on the social network involves both positive and negative relationships, e.g., since attempts to establish a relationship may fail due to decline from the expected target. This generates links that signify rejection of invitations, disapproval of applications, or expression of disagreement with others' opinions. Such social networks are mutual since the sign of a link indicating whether it is positive or negative depends on the attitudes or belief of both entities forming the link. Moreover, mutual positive and negative relationships have been even less investigated.
Fig.1: Collaborative Information for Composite and Mutual Link Prediction
Interaction Predicted Preference Predicted match
Recently, social network analysis has had a variety of applications, such as online dating sites, education admission portals as well as jobs, employment, career and hiring sites, where people in the networks have different roles and links between them can only between people in different roles.
Social networking is a concept that has been around for a long time than the Internet. People have always been able to work together in a team. Social networking has come to users using the internet to communicate in various different ways. This focuses on creating a ground for social networking and collaboration. Social networking is about everyone in the society where it has become self-sustaining and created further growth leading to human social interaction. A social networking site has allowed users to post their profiles and create personal details for exchanging information with other users (Weaver and Morrison, 2008) [1]. In early stages, social networks have become the most frequently used tools. “Facebook is currently the largest online social network. Its business model is based on the analysis of user data to display customized advertisements. However, the data collection induces possible privacy concerns which oppose perceived benefits. Information privacy concerns are important aspects for the intention not to use Facebook but it is outweighed by perceived usefulness”. (Becker and Pousttchi 2012, pg.187) [2].
Society is increasingly subjected to predictions on subjects as diverse as economic development, finance, fashion and even relationships. For instance, Economists forecast the gross domestic product of countries; Financial Analysts model the likely increase in earnings per share of a company based on potential sales of future products; Fashion forecasters predict how the mood of consumers determine the styles for next season’s haute couture collections; and websites encourage a person to input data about them self and an algorithm tries to predict their most suitable partner.
Social media are internet-based information tools and technologies used to share information and facilitate communications between internal and external audiences. There are various well-known social media platforms, such as LinkedIn, Facebook, and Twitter that come to mind when talking about social media in recruitments. According to the 2013 SHRM research, these social networks are the most commonly used for recruitment, with the vast majority of organization, about 94%, using LinkedIn, followed by Facebook (54%), twitter (39%) and other professional social network site other then SHRM Connect (29%)
33 Barabási, A. L., 2002 Linked: The New Science of Networks, Cambridge, MA, Perseus Publishing pp 8-9
The Social networking sites have full-grown in popularity over the last few years, mostly between teenagers and young adults. Also they are web-based services that they will first ask about the register and after that the users can become a memberships in the sites. The customers will do this by filling out the working forms and it will include their age, their hobbies and their location. They can also upload their personal photos, and they can modification their profile looks also they can add multimedia content. And after that the users can communicate with each other. There is another social network site it’s the SixDegrees.com and it was the first site recognized as social network. This site allowed all the users to create their own profiles and they can list their friends also it allowed for surfing of friends. And between 1997 to 2001 there are several of sites that have the features of the social networking. Moreover, there are a lot of new social network site were hurled. And most of them try to repeat Friendster's early achievement or targeted precise demographics. MySpace was launched in 2003 to attract Friendster that wants to adopt a fee-based system and it grew so quickly and it welcomes the groups that join in. In 2004 Facebook has started as a Harvard-only social networking site.
Social network, also referred to as a virtual community, is a platform that unites individuals to communicate. For instance, using it to email co-workers or instant message friends. Although it makes people's lives easier, it serves as an adverse tool of communication. The individual won't have face to
Based on the number of ties a node has, its reach to all other nodes, and its position in controlling the flow between other nodes, Freeman (1978) classified three measures of centrality: degree, closeness, and betweenness. The weighted degree of the node is the sum of the weights attached to the ties connected to a node (Barrat et al., 2004). Betweenness centrality is based on “the identification of the shortest paths, and measures the number of them that passes through a node” (Opsahl et al. 2010, p. 247). In a weighted network, the quickest path may be through nodes that are stronger and have more frequent ties than through a weak direct tie (see Opsahl et al., 2010). Closeness centrality refers to the average distance from a node to all other nodes in the network. The closeness centrality measure is somewhat limited in its applicability when there are disconnected components in our network and is restricted in its use within the giant component (see Opsahl et al, 2010). Eigenvector centrality, defined as the “weighted sum of not only direct connections but indirect connections of every length” (Bonacich, 2007, p. 555) is another important measure used in the exploration of our network that allowed us to identify individuals who are connected to other well-connected individuals. In order to measure local
The use of social networking has both its advantages and disadvantages. One advantage is “an Internet social network can help you connect with other people who share your interest, and find resources to ga...
Twitter, Skype, Facebook these are just a few of the online social networks we utilize day to day, which has made connecting to others easier than before. A social network is a structure made up of individuals or organizations that are tied by one or more specific types of relationships such as friendships. Although traditionally operated with person to person contact, it is now more popular online through social media networks such as Facebook and Skype. There are millions of persons with wide ranges of personalities who are looking to develop new friendships or to simply become a part of a group in order to share information on these websites. As of July 2011, Facebook which continues to be the most popular social network had 750 million active users compared to February 2009 when it was at a mere 175 million. Given the ubiquitous presence of online social networks, it is not surprising that they have revolutionized society through communication which provides fast access to information, providing a platform for socialization with different cultures then transforming the world into a global village.
Social networks are increasing dramatically every year. Employers are turning to social networks because it is a tool to screen job applicant’s profiles. According to a survey conducted by jobvite.com (2013), 94 % of employers use social media profiles to recruit job applicants. This trend assists the applicants and recruiters. Job applicants should be judged by their social network profiles because social media give positive image about the candidate, prove the information in the resume, and help to identify if the person fits the culture of the company or not.
One of the ways Social Network Theory has been used is to examine how companies interact with each other. By understanding what links members of the companies together, Social Network theory has served as a way to gather information about the relationships within their company to help answer questions through the roles of the relationships within the company. For my question of study I looked at the question of: Are actors and their actions viewed as independent of interdependent within an organization?
Internet has become a vital element in people daily lifestyles. People use smart phones, tablets, laptops or computers to access Internet. By the first decade of the 21st century, many Internet users use faster broadband Internet access technologies. As the Internet users grow, one of the Internet phenomenons that can be seen is social networking. Basically, people use social media to interact among people where they create, share or exchange information in virtual communities and networks.
The growth of social networking is one of the fastest growing digital trends to exist. Many social networking sites boast with millions if not billions of members. Prominent examples of these social sites are Facebook, Twitter, and LinkedIn. Members of these networks use them daily to communicate, share various types of information or to collaborate with other members.
Many social networks will tell you that the more you put into them (your information), the more you get out of them (connections, recommendations, etc.). However, despite the aura of privacy they try to engender, one must keep in mind that social network takes place in essentially public space, with only the barest of mechanisms providing any semblance of privacy. Even seemingly innocuous data shared with the world can be dangerous in the wrong hands.
Network analysis has been adopted across the scientific spectrum from the social sciences to biochemisty with applications in empirical research, modelling, and management, to name a few.1,2,3,4 While the network structure of operating sub-groups has been examined previously to our knowledge a comprehensive analysis of the operating suite incorporating all relevant participants has not yet occurred.5 In studying a network several definitions are worth reviewing (Table 1). Networks can be directed or undirected, referring to whether an edge has a defined source and target or merely denotes the existence of a connection, and weighted or unweighted, referring to value attributed to an edge to impart information related to the nature of a tie. From the structure that arises out of nodes and their corresponding edges several traits addressing their importance in a network can be discussed.9 Though many and varied measures exist, perhaps the most commonly discussed centralities are degree, closeness, betweenness, and eigenvector and it is to these we will limit our analysis. Further, the empirical detection of existing sub-groups or “communities” in a the network will also be examined. While clustering coefficient, a common metric of the social embeddedness of a node in a network, can be applied to undirected networks it cannot be applied to weighted networks and thus was not examined.10