MUTUALLY EXCITING POINT PROCESSES FOR CROWDFUNDINGPLATFORM DYNAMICS
成果类型：
Article
署名作者：
Djorno, Alexandra; Crawford, Andforrest W.
署名单位：
Yale University
刊物名称：
ANNALS OF APPLIED STATISTICS
ISSN/ISSBN：
1932-6157; 1941-7330
DOI：
10.1214/25-AOAS2051
发表日期：
2025-09
页码：
1931-1947
关键词：
摘要：
Crowdfunding is a powerful tool for individuals or organizations seekingfinancial support from a vast audience. Despite widespread adoption, man-agers often lack information about dynamics of their platforms. Hawkes pro-cesses have been used to represent self-exciting phenomenon in a wide vari-ety of empirical fields but have not been applied to crowdfunding platformsin a way that could help managers understand user behaviors. In this paperwe extend the Hawkes process to capture important features of crowdfundingcontributions and apply the model to analyze data from two donation-basedplatforms. For each user-item pair, the continuous-time conditional intensityis modeled as the superposition of a self-exciting baseline rate and a mutualexcitation by preferential attachment, both depending on prior user engage-ment and attenuated by a power law decay of user interest. The model isthus structured around two time-varying features, contribution count and itempopularity. We estimate parameters that govern the dynamics of contributionsfrom 2000 items and 164,000 users over several years. We identify a bottle-neck in the user contribution pipeline, measure the force of item popularity,and characterize the decline in user interest over time. A contagion effect isintroduced to assess the impact of item popularity on contribution rates. Thismechanistic model lays the groundwork for enhanced crowdfunding platformmonitoring based on evaluation of counterfactual scenarios and formulationof dynamics-aware recommendations
来源URL：