In this paper we address the issue of how users’ personality affects the way people interact and communicate in Facebook. Due to the strict privacy policy and the lack of a public timeline in Facebook, we automatically sampled data from the timeline of one “access user”. Exploiting Facebook’s graph APIs, we collected a corpus of about 1100 ego-networks of Italian users (about 5200 posts) and the users that commented their posts. We considered the communicative exchanges, rather than friendships, as a network. We annotated users’ personality by means of our personality recognition system, that makes use of correlations between written text and the Big5 personality traits, namely: extroversion, emotional stability, agreeableness, conscientiousness, openness. We tested the performance of the system on a small gold standard test set, containing statuses of 23 Facebook users who took the Big5 personality test. Results showed that the system has a average f-measure of .628 (computed over all the five personality traits), which is in line with the state of the art in personality recognition from text. The analysis of the network, that has a average path length of 6.635 and a diameter of 14, showed that open-minded users have the highest number of interactions (highest edge weight values) and tend to be influential (they have the highest degree centrality scores), while users with low agreeableness tend to participate in many conversations.

Relationships between personality and interactions in facebook

Luca Polonio
2013-01-01

Abstract

In this paper we address the issue of how users’ personality affects the way people interact and communicate in Facebook. Due to the strict privacy policy and the lack of a public timeline in Facebook, we automatically sampled data from the timeline of one “access user”. Exploiting Facebook’s graph APIs, we collected a corpus of about 1100 ego-networks of Italian users (about 5200 posts) and the users that commented their posts. We considered the communicative exchanges, rather than friendships, as a network. We annotated users’ personality by means of our personality recognition system, that makes use of correlations between written text and the Big5 personality traits, namely: extroversion, emotional stability, agreeableness, conscientiousness, openness. We tested the performance of the system on a small gold standard test set, containing statuses of 23 Facebook users who took the Big5 personality test. Results showed that the system has a average f-measure of .628 (computed over all the five personality traits), which is in line with the state of the art in personality recognition from text. The analysis of the network, that has a average path length of 6.635 and a diameter of 14, showed that open-minded users have the highest number of interactions (highest edge weight values) and tend to be influential (they have the highest degree centrality scores), while users with low agreeableness tend to participate in many conversations.
2013
978-1-62808-534-1
Social Network Analysis, Personality Recognition, Facebook, Natural Language Processing.
File in questo prodotto:
File Dimensione Formato  
Chapter_social Networking(2013).pdf

accesso aperto

Descrizione: Articolo principale
Tipologia: Documento in Post-print
Licenza: Nessuna licenza
Dimensione 238.11 kB
Formato Adobe PDF
238.11 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11771/12524
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
social impact