IJSRD - International Journal for Scientific Research & Development| Vol. 3, Issue 08, 2015 | ISSN (online): 2321-0613 All rights reserved by www.ijsrd.com 548 A Recommender System for Online Social Network based Persona Analysis Prof. Ranjeetsingh Suryawanshi 1 Ankit Sharma 2 Ajay Singh Rautela 3 Akshay Trimbake 4 Sanket Vangari 5 1 Professor 2,3,4,5 Student 1,2,3,4,5 Department of Computer Science 1,2,3,4,5 Trinity College of Engineering and Research, Pune Abstract—Our primary goal is to design a recommender system for analyzing the personality traits of people, in online social networks, for hiring potential candidates via an online recruiting system where the social and cultural aspects of a client will be assessed by the HRMs using data mined from posts, comments, etc. on/by the clients on social media. Key words: Recommender System, Persona Analysis I. INTRODUCTION Human Resource Management (HRM) has recently shown an upcoming inflow of networking and network-based devices [1]. In this paper we primarily focus on designing a system that points out the personality traits of a Candidate for Human Resource Management activity. II. EXISTING WORK Currently existing systems have their focus mainly on the concerns of providing a candidate who fits the job description based on the skills defined by the candidates in their resumes, CVs, etc.[1].None of these systems can ensure the cultural fitness of the candidate with respect to the hiring company’s social culture. This happens because that would require a thorough knowledge of the candidate’s personality traits that can only be measured when some time is spent with the candidate. Over the past 10 years, social interactions using Online Social Networking Sites like Facebook and Twitter have become unavoidable. Regardless of their psychological productivity, OSNs have remained relatively unexplored from an experiential outlook. This can be better explained by going through the Case Studies carried out on:  Personality and Self-Reported Facebook-Related Behaviors.  Personality and Observable Information on Facebook Profiles. Currently, the most popular and widely used Online Social Networking site in the world is Facebook. The data on Facebook has been studied thoroughly, mapping out the connection between the OSN behavior and personality traits [3][4]. The objective of Study 1 was to examine the affiliation between personality and a domain of self-reported Facebook activities. It was found that people who spent more time socializing in the offline world also tend to spend more time socializing online [5]. People were assessed on the five factor model (FFM) which consists of personality dimensions (agreeableness, neuroticism, extraversion, conscientiousness, and openness to experience) [6]. Extraverts are people who seek other people’s attention and make their presence evident. The number of Facebook friends and comments on various pages show the extrovert nature. Agreeableness is higher in people who viewed or visited more pages compared to people who visited less number of pages. People who spent more time viewing pages are low on conscientiousness. People with open nature (Openness) tend to upload or replace more photographs. No relations were found between Neuroticism and any of the self-reported Facebook behaviors [3]. In Study 2, the relation between the observable information found on Facebook profiles and personality traits are studied [4]. Extraversion and Openness are the only dimensions that were related to the information extracted from Facebook. The overall number of friends on Facebook unveiled the extravert nature of the individual and the number of friends in the limited or local network exhibited the openness [3]. Except Extraversion and Openness, all other dimensions were not related to the observable Facebook information. An analysis carried on establishing of connections through social networks, tell us that social networks are made in a way that they support efficient local/global search and communication. Also, while giving this a thought, one might get ideas that might be applicable to social networking sites such as LinkedIn, Facebook, Myspace, etc., that are being used all over the globe. The establishment of these conclusions is derived by studying a survey on basic positivity and negativity observed in posts and comments made on Online Social Networks as follows: A. Initial Data Search and Collection Survey The Survey started with the most popular Social Networking Sites (SNS), where initially focus was on number of wall posts. The chart shows the number of wall posts per forum of a random survey: Forum Name Total wall posts Positiv e Posts Negativ e Posts Survey Year Scientific American 689,00 3 47% 22% Januar y 2012 Reference paper[9] 90,269 66.7% 25.3% Januar y 2009 pewresearch.or g 122 million words 4 million 1.8 million July 2014 Table 1: Survey on Positive and Negative Posts [9]. After this, a search of words was performed, using both, positive keywords and negative keywords which led to the generation of posts from the keywords in the database [10].