Network Regression and Supervised Centrality Estimation

Cai, Junhui; Yang, Dan; Zhu, Wanpeng; Shen, Haipeng; Zhao, Linda

doi:10.2139/ssrn.3963523

Cited by 1 publication

(2 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Second, the social network data contains inherent measurement errors in various contexts [29][30][31][32]. In particular, when network data is sparse, the ordinary least squares estimator might be inconsistent, as recently explored by econometric research [29]. Due to these limitations, the generalizability of our findings may be restricted, and hence, a caveat to our conclusions is warranted.…”

Section: Plos Onementioning

confidence: 98%

“…In spite of their external selection by Bharatha Swamukti Samsthe, a microfinance institution in rural India, we do not interpret our result as indicating causality between diffusion rates and centrality measurements. Second, the social network data contains inherent measurement errors in various contexts [29][30][31][32]. In particular, when network data is sparse, the ordinary least squares estimator might be inconsistent, as recently explored by econometric research [29].…”

Section: Plos Onementioning

confidence: 99%

See 1 more Smart Citation

Who are the key players? Listeners vs spreaders vs others

2023

View full text Add to dashboard Cite

The literature on social learning examines how people learn from their neighbors and reach a consensus. The DeGroot social learning model describes the social learning process as one in which people form their opinions by taking a weighted average of their neighbors’ opinions. In the model, the influence structure is represented by a single matrix. In this paper, we empirically identify the role of the hub and authority centralities based on this matrix using data on microfinance adoption in rural Indian villages. Controlling for other well-known centrality measures, authority centrality is positively associated with final adoption rates in the villages, but hub centrality is not. Furthermore, we find that authority centrality is the most informative variable predicting microfinance diffusion success from LASSO regressions.

show abstract