Power-law distributions in empirical data
TL;DR
Maximum likelihood estimation for power law exponents
Speed Read
0%
300 WPM
-- remaining
Definitive statistical framework for detecting power laws in real data. Provides maximum likelihood estimation methods, goodness-of-fit testing, and likelihood ratio tests comparing power laws to alternative distributions. Essential reading for anyone claiming to observe power laws.
The paper demonstrates that many purported power laws in literature don't survive rigorous testing, establishing higher standards for power law claims. Over 9,400 citations make this the authoritative reference for power law methodology.
Key Findings from Clauset et al. (2009)
- Maximum likelihood estimation for power law exponents
- Goodness-of-fit testing via Kolmogorov-Smirnov statistic
- Likelihood ratio tests vs. log-normal, exponential alternatives
- Many claimed power laws fail rigorous testing
- Software implementations provided
Cited in 8 pages
Mechanism Power Law Distributions Mechanism Scale Invariance Framework Power Law Strategy Framework Framework Power Law Detection Methods Citation A brief history of generative models for power law and lognormal distributions Citation Power laws, Pareto distributions and Zipf's law Citation Cours d'économie politique Citation The Black Swan: The Impact of the Highly Improbable