Modelling the loss and retention of contacts in social networks: The role of dyad-level heterogeneity and tie strength

被引:4
|
作者
Calastri, Chiara [1 ,2 ]
Hess, Stephane [1 ,2 ]
Daly, Andrew [1 ,2 ]
Antonio Carrasco, Juan [3 ]
Choudhury, Charisma [1 ,2 ]
机构
[1] Univ Leeds, Inst Transport Studies, Leeds, W Yorkshire, England
[2] Univ Leeds, Choice Modelling Ctr, Leeds, W Yorkshire, England
[3] Univ Concepcion, Concepcion, Chile
基金
欧洲研究理事会;
关键词
Hybrid choice models; Social networks; Random taste heterogeneity; Intra-respondent; TRAVEL BEHAVIOR; PERSONAL NETWORKS; DISCRETE-CHOICE; DYNAMICS; MEMBERS; DECISION; EXAMPLE; LIFE;
D O I
10.1016/j.jocm.2018.03.004
中图分类号
F [经济];
学科分类号
02 ;
摘要
Social networks have attracted attention in different fields of research in recent years and choice modellers have engaged with their analysis by looking a the role that social networks play in shaping decisions across a variety of contexts. The incorporation of the social dimension in choice models creates the need for understanding how social networks evolve over time and in particular which social contacts (alters) are retained over time by an individual (ego). Existing work fails to capture the full extent of ego-level and ego-alter level heterogeneity in these processes. We propose the use of a hybrid model framework which is based on the notion of latent strength of relationship. The resulting model allows for heterogeneity in the latent strength both across individuals and across their different relationships. In addition, we allow for heterogeneity not linked to the latent strength concept. We demonstrate the benefits of the approach using data from Chile, showing the presence of extensive variations in retention of social contacts and in strength of relationship both at the ego and ego-alter level, only some of which can be linked to observed characteristics.
引用
收藏
页码:63 / 77
页数:15
相关论文
共 1 条
  • [1] Disentangling degree and tie strength heterogeneity in egocentric social networks
    Heydari, Sara
    Iniguez, Gerardo
    Kertesz, Janos
    Saramaki, Jari
    EPJ DATA SCIENCE, 2024, 13 (01)