Inferring Online Relationships from User Characteristics

• Main Authors: Suwimon VONGSINGTHONG, Sirapat BOONKRONG, Herwig UNGER
• Format: Article
• Language: English
• Published: Walailak University 2018-04-01
• Series: Walailak Journal of Science and Technology
• Subjects: Diffusion; stochastic model; structural evolution; temporal behavior; user characteristics
• Online Access: http://wjst.wu.ac.th/index.php/wjst/article/view/3541

A new approach to understanding human online behavior in regard to psychological functioning is proposed through the developed user’s activity model incorporating the influences of social behavior, network, and content. Microscopic levels of user characteristics induced by personality traits were interpolated as interaction rules, whilst an unsupervised clustering algorithm was applied to penetrate the individual complexity. Temporal behavior of disparate users was mimicked, and streaming network data was generated and computationally analyzed. A comprehensive understanding of how individuality, friendship, and varying temperaments dramatically reshaped the networks was gained from insight synthesis of network properties characterized by small-world, scale-free, and centrality measures. Evidence illustrates that users with high extraversion possess high numbers of friends and spread massive information, while high conscientious and high intellect users are seriously discreet in accepting friends and often produce influential content. These results not only provide a wealth of challenges for product recommendation, network structure optimization, and design, but also are useful for the prediction of future network structural evolution.