1 Revealing Psycholinguistic Dimensions of Communities in Social Networks Tushar Maheshwari, Aishwarya N. Reganti, Upendra Kumar, Tanmoy Chakraborty 1 , Amitava Das {tushar.m14,aishwarya.r14,upendra.k14,amitava.das}@iiits.in, 1 tanmoy@iiitd.ac.in Indian Institute Of Information Technology Chittoor, Sri City, A.P. , India 1 Indraprastha Institute of Information Technology Delhi (IIIT-D), India Abstract: A community in social network is composed of individuals with similar behavior. Although there has been a plethora of work on understanding network topologies (edge density, clustering coefficient, etc.) within a community, the semantic interpretation of a community has hardly been studied. The present paper aims at understanding Personalities and Values of individuals in social communities. To this end, we collect datasets from various social media platforms (including Facebook, Twitter), which contain Values and Ethics of users. Then we design a three-fold experimental setup. First, we propose automatic models to determine Personality and Values (Values and Ethics) of individuals by analyzing the kind of language used and their conduct in social media. Secondly, various experiments are performed to understand the blend of characteristics of individuals within a social network community. Finally, we claim that the detected Personality and Values of individuals can be used further as additional node attributes to detect better community structure. To our knowledge, this is the first computational analysis to understand and predict psycholinguistic dimensions of individuals in social networks. Index Terms—Personality, Values, social media, societal sentiment, community ✦ 1 I NTRODUCTION D ETECTING and analyzing dense groups or commu- nities from social networks has attracted immense attention over the last decade. Several heuristics and al- gorithms were proposed to detect communities based on the topological structure of the network [1]. However, the semantic interpretation of a community, i.e., the behavior of individuals within a community has hardly been studied. This paper presents the first computational psycholinguistic study to understand the behavior of individuals forming communities in social networks. We use two psycholinguis- tic models – Personality and Values models, to identify the behavior of individuals. The Personality model is used to understand the characteristics or blend of characteristics at individual level, whereas the Values model is used to analyze inter-personal dynamics of societal sentiment. The Big 5 Personality traits [2], aka the five factor model (FFM), is a widely used Personality model. The five factors are: Openness (O): A personality trait possessed by in- dividuals who are imaginative, insightful and have wide interests; Conscientiousness (C): Refers to those who are organized, thorough, and planned; Extroversion (E): Refers to Personality of those who are talkative, energetic, and assertive; Agreeableness (A): Individuals with this Per- sonality trait are sympathetic, kind, and affectionate; and Neuroticism (N): Individuals who are mostly tense, moody, and anxious. Big 5 model is also represented using the acronym OCEAN. To define societal sentiment, we use the well-established “Schwartz Theory of Basic Human Values” [3], which de- fines ten basic and distinct personal Values – Achievement (AC): Achievers sett goals and then make all possible en- deavors to achieve them; Benevolence (BE): Those who tend towards being benevolent are very philanthropic, they seek to help others and provide general welfare; Confor- mity (CO): This category of people obey clear rules and structures; Hedonism (HE): Hedonists are those who simply enjoy themselves; Power (PO): The ability to control others is important to people who possess this value and power will be actively sought by dominating others and control over resources; Security (SE): Those who seek security value, health and safety to a greater extent than other people (perhaps because of childhood woes); Self-direction (SD): Individuals who are self-directed, enjoy being independent and are outside the control of others; Stimulation (ST): It is closely related to hedonism, nevertheless the goals are slightly different. In this case, pleasure is acquired specifically from excitement and thrill; Tradition (TR): A traditionalist respects practices of the past, doing things blindly because they are customary; Universalism (UN): Individuals who are universal, seek social justice and tol- erance for all. Shwartz, along with the identification ten basic Values, also explains how these Values are related to each other and influence one another, since individuals possessing any of the Values may also possess values in accord with another (e.g., Conformity and Tradition) or contrary with at least one other Value (e.g., Benevolence and Achievement). Such coinciding nature psychological classes makes the compu- tational classification problem much more challenging than the typical sentiment analysis problem. In this paper, we raise a fundamental question in order to understand the behavioural characteristics of individuals This article has been accepted for publication in IEEE Intelligent Systems but has not yet been fully edited. Some content may change prior to final publication. Digital Object Identifier 10.1109/MIS.2018.111144400 1541-1672/$26.00 2018 IEEE