Over the last few decades power law distributions have been suggested asforming generative mechanisms in a variety of disparate fields, such as,astrophysics, criminology and database curation. However, fitting these heavytailed distributions requires care, especially since the power law behaviourmay only be present in the distributional tail. Current state of the artmethods for fitting these models rely on estimating the cut-off parameter$x_{\min}$. This results in the majority of collected data being discarded.This paper provides an alternative, principled approached for fitting heavytailed distributions. By directly modelling the deviation from the power lawdistribution, we can fit and compare a variety of competing models in a singleunified framework.
展开▼
机译：在过去的几十年中，已经提出幂律分布是形成各种不同领域的生成机制，例如天体物理学，犯罪学和数据库管理。但是，拟合这些繁重的分布需要格外小心，尤其是因为幂律行为可能仅出现在分布尾部。拟合这些模型的最新方法依赖于估计截止参数$ x _ {\ min} $。这导致大多数收集的数据被丢弃。本文提供了一种适合粗尾分布的替代原则方法。通过直接建模与幂律分布的偏差，我们可以在一个统一的框架中拟合和比较各种竞争模型。
展开▼
