Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Pareto principle index

12 views
Skip to first unread message

avi

unread,
Apr 25, 2013, 6:15:24 PM4/25/13
to
The pareto principle states that about 20% of the possible causes account for about 80% of the occurrences (for example, 20% of the possible defect are responsible to 80% of the total defective units)

I wonder if there is some general calculated index to indicate the extent to which the principle holds for a certain empirical distribution, regardless the field in question

With such an index , I will be able for example to compare many sub populations regarding the extent to which they follow the pareto principle

I'm aware of the GINI index in economics, but looks for a more universal method

Thanks
Avi

Rich Ulrich

unread,
Apr 27, 2013, 4:24:12 PM4/27/13
to
On Thu, 25 Apr 2013 15:15:24 -0700 (PDT), avi <avi...@bezeqint.net.il>
wrote:

>The pareto principle states that about 20% of the possible causes account for about 80% of the occurrences (for example, 20% of the possible defect are responsible to 80% of the total defective units)
>
>I wonder if there is some general calculated index to indicate the extent to which the principle holds for a certain empirical distribution, regardless the field in question

When I read this closely, I am more confused than
when I read it hastily.

Are you trying to figure out when "20-80/80-20" fits,
as a particular set of numbers? That is the most used
classical example, I guess, but there is nothing magical
about it.

Or are you trying to say whether whole principal is appropriate?


>
>With such an index , I will be able for example to compare many sub populations regarding the extent to which they follow the pareto principle
>
>I'm aware of the GINI index in economics, but looks for a more universal method
>

The Wikip artilcle on the Gini coefficient is pretty good.
It shows a nice graph of what is being measured.
It also mentions the mathematical generalization for
a general formula where weights can be varied.

The Gini index measures the area under the curve. There
is a section in the article that discusses when and how it
can be misleading to make comparisons of Gini indexes.

A minimum of consideration shows that simple comparisons
are made with least ambiguity when the curve is smooth
with the appropriate symmetry.

- I suppose it might be interesting to have an extra
number to describe, "How smooth and symmetrical
is the curve" -- to indicate how misleading Gini might be
for a particular case.

Did you have something else in mind?

--
Rich Ulrich
0 new messages