Re: How can I order a aggregation of items from multiple lists
- From: Richard Ulrich <Rich.Ulrich@xxxxxxxxxxx>
- Date: Fri, 22 Dec 2006 23:45:35 -0500
On 22 Dec 2006 11:14:20 -0800, "bcot" <arustad@xxxxxxxxx> wrote:
I hope this is an appropriate groups for asking this question.
I'm trying to calculate what an aggregate list of items should look
given multiple small lists of items. Items may appear in in multiple
lists. And each list is ordered from most popular to least popular.
For example:
user 1 list user 2 list user 3 list
-------------- --------------- ---------------
a a b
b b x
c e y
d f
e
f
Given the data above, should 'b' be the first time, because it shows up
in 3 lists? Or, because there are 2 'a' and they are both in position
1, should they be worth more?
Any help on this would be appreciated...obviously, I'm no mathmatician,
so if I need to provide further info, please let me know.
When it comes to decisions here,
the mathematics take second place to the politics.
Do you want to maximize the *satisfaction* of a few raters,
or to minimize the *dissatisfaction* across all raters?
If you score this like Victories in a track meet or a swimming
meet between schools, individual events are given points.
A "1" might count 10 points for the team, a "2" count 5 points,
and places "3", "4", and "5" count 3, 2, and 1, respectively.
Relay races count double.
Dealing with ranks, you get into further choices if you want
to rank the whole set -- like voting for ice skaters in
Olympic competition (by the old, complicated system, or
the new, complicated system) -- or if you want to choose one
winner for a political office with a dozen parties in the race.
(Various places do have various systems.)
In general, the practice is for you to do the following -
1) Look at a *lot* of examples of what you are rating,
including a number of extreme cases.
2) Figure out who you *feel* should be the winners.
3) Find a system that gives you the winners that you
feel are right, with the least work and confusion.
That's trickier if the "you" is plural.
What's easier is to copy pretty closely whatever other
folks have done with *your* exact problem ....
--
Rich Ulrich, wpilib@xxxxxxxx
http://www.pitt.edu/~wpilib/index.html
.
- References:
- Prev by Date: Re: meaning of "the effect is significant at the level of .000000004"?
- Next by Date: Re: Matrix derivative
- Previous by thread: How can I order a aggregation of items from multiple lists
- Next by thread: Is the Monty Hall Problem like Quantum Mechanics?
- Index(es):
Relevant Pages
|
