clauset.xmin {distributionsrd} | R Documentation |
This method determines the optimal scale parameter of the Pareto distribution using the iterative method (Clauset et al. 2009)that minimizes the Kolmogorov-Smirnov distance.
clauset.xmin(x, q = 0)
x |
data vector |
q |
Percentage of data to search over (starting from the largest values) |
Returns a named list containing a
Named vector of coefficients
Minimum Kolmogorov-Smirnov distance
Number of observations in the Pareto tail
Evolution of the Pareto shape parameter over the iterations
Clauset A, Shalizi CR, Newman ME (2009). “Power-law distributions in empirical data.” SIAM review, 51(4), 661–703.
## Determine cuttof from compostie lognormal-Pareto distribution using Clauset's method dist <- c("lnorm", "pareto") coeff <- c(coeff1.meanlog = -0.5, coeff1.sdlog = 0.5, coeff2.k = 1.5) x <- rcomposite(1e3, dist = dist, coeff = coeff) out <- clauset.xmin(x = x) out$coefficients coeffcomposite(dist = dist, coeff = coeff, startc = c(1, 1))$coeff2 ## Speed up method by considering values above certain quantile only dist <- c("lnorm", "pareto") coeff <- c(coeff1.meanlog = -0.5, coeff1.sdlog = 0.5, coeff2.k = 1.5) x <- rcomposite(1e3, dist = dist, coeff = coeff) out <- clauset.xmin(x = x, q = 0.5) out$coefficients coeffcomposite(dist = dist, coeff = coeff, startc = c(1, 1))$coeff2