Talk:Pareto interpolation

From Wikipedia, the free encyclopedia

This article is within the scope of WikiProject Statistics, which collaborates to improve Wikipedia's coverage of statistics. If you would like to participate, please visit the project page.

It looks as if the formulas given here find the median of the distribution, given the parameters, rather than the median of a "set of data", as advertised in the first paragraph. A "set of data" would be a finite list of numbers, on which estimates of the parameters would be based, including an estimate of the population media. The article doesn't say how to find that estimate given a finite list of observations. That seems to be promised by the opening sentence. Could someone add that? Michael Hardy 00:51 Mar 28, 2003 (UTC)

[edit] sigh....

It's been more than three years, and no one's fixed this. I don't know the topic, but I'm tempted to edit according to my surmises.

  • "to find the median of a set of data" makes no sense! Here is my surmise: It is intended to estimate the median of a population from which the data were taken as a sample. Obviously the proposed method will not find the median of the data!
  • I surmise that κ is intended to be the lower bound of the support of the distribution and θ the Pareto index. If so, the article should say so!!
  • What are the "categories" referred to??????? My surmise is that either the data are expressed as ranges---thus "19% of the individuals in the sample had incomes between $28000 per year and $33000 per year" etc., or else that the person using this technique is expected to put the data into that form first---in which case it wouldn't hurt to at least mention the issue of where to draw the boundaries. Either way, the "categories" need to get explained!
  • It looks to me as if a non-statistican learned this in an economics course and doesn't know how to think about these sorts of things, and wrote this article.

Michael Hardy 23:21, 15 April 2006 (UTC)

...now I've looked on both Google Scholar and the Current Index to Statistics data base and I cannot find "Pareto Interpolation". A reference would help! Michael Hardy 23:29, 15 April 2006 (UTC)

... and now I've worked through the math carefully and it has become clear what must have been intended. I'll be back Monday, when I'll probably re-write this article from scratch. Michael Hardy 02:17, 16 April 2006 (UTC)

OK, I've substantially re-written this thing. It could still use some references. Michael Hardy 23:15, 16 April 2006 (UTC)