Subject:
|
Re: Voting algorithm?
|
Newsgroups:
|
lugnet.off-topic.geek
|
Date:
|
Sun, 24 Dec 2000 18:14:27 GMT
|
Viewed:
|
63 times
|
| |
|
|
In lugnet.off-topic.geek, Christopher Lindsey writes:
> I have a Web site with about 1400 of my plant photos. I recently added
> a voting mechanism to help me order images of the same plant by quality
> and to help me figure out which need to be replaced.
>
> The problem that I'm seeing is that people aren't voting consistently.
> In some cases an image will garner a vote of 5 out of 5, but someone
> else will give the same image a 1 out of 1.
>
> LUGNET gets around this by displaying all votes as a stem and leaf graph,
Well, actually, it's a standard histogram, with snapping to tens on the range.
In contrast, a stemplot (a.k.a. stem and leaf graph) uses the numeric digits
of the constituent values as part of the actual bars, which is a bit different.
But they do look rather similar.
> but I'm trying to figure out how to order images based on their ratings.
> Has anyone else done this? I think I need to go with the median or mode
> instead of the mean, but I'm wondering if anyone has any real-world
> experiences (positive or negative) to share about this.
How about using the mean for the rating, but with a confidence factor based on
some combination of the standard deviation and the number of votes? Or, if
you aren't concerned about the variance, you could include an implicit vote of
3 as a dampener in all cases and use purely the mean -- that automatically
factors in the number of votes and prevents pegging.
--Todd
|
|
Message is in Reply To:
 | | Voting algorithm?
|
| I have a Web site with about 1400 of my plant photos. I recently added a voting mechanism to help me order images of the same plant by quality and to help me figure out which need to be replaced. The problem that I'm seeing is that people aren't (...) (24 years ago, 24-Dec-00, to lugnet.off-topic.geek)
|
2 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
