ABSTRACT
In the majority of practical applications of the Pareto distribution, sample sizes are likely to equal or exceed this magnitude. A 100(1 - 'r))% confidence interval on & (at least in the case of complete samples) can be calculated from the chisquare distribution as
2n 2n (10.4.3)
For large values of N, we might take advantage of the approach to normality to approximate a 95% CI as
-y~ + 1.96 I . (10.4.5) N&- This same confidence interval should also provide a satisfactory approximation when it is based on modified maximum likelihood estimates.
