Specificity calculation
Moderators: Stardust@home Team, DustMods
Specificity calculation
My Specificity value remains at 100% while the Sensitivity bounces around. I think something is amiss.
-
- Posts: 3
- Joined: Thu Sep 13, 2007 11:20 am
I'm a little bit grumpy right now about my ratings, which it seems are down in the pits! Is there anybody besides me who is curious about the average Sensitivity and Specificity ratings, and how they compare between Phase 1 and Phase 2? Maybe comparing the Top 100 vs. overall?? Seems it would be a simple query of the user database to find out...
Or maybe it would make me even more depressed about my ratings....
Or maybe it would make me even more depressed about my ratings....
ratings
I'd like to know other people's ratings too. I'm uncomfortable about how many mistakes I make.
My ratings are Spec. at 96%. Sensitivity is only 84%, yet I seem to identify a lot of tracks that other people are not identifying as tracks.
Hopefully some other people will share their scores.
Dee
My ratings are Spec. at 96%. Sensitivity is only 84%, yet I seem to identify a lot of tracks that other people are not identifying as tracks.
Hopefully some other people will share their scores.
Dee
Out of 220 CMs, my specificity is still 100%, but the sensitivity 82%. We can't complain: they warned us that the rates will go down! And I try to keep in mind, that I'm not looking for CM-points, but for possible tracks in the actual Stardust-movies. Spotting those tracks seem a lot tricker than we were (in my opinion) led to believe: Many of the passed candidate tracks are such as I wouldn't have dreamed of clicking on during Phase I! But that's science.
-
- Stardust@home Team
- Posts: 530
- Joined: Wed Apr 19, 2006 11:28 am
- Location: UC Berkeley Space Sciences Lab
- Contact:
Indeed that is science. We are also surprised at the subtle nature of the best candidate tracks.
I could do a search on the database to figure the average sensitivity, however right now I'm having difficulty connecting. I'll try again later.
If the average is down about 80% then that means we at about the level where our smallest CM tracks are at the detection threshold for our set of volunteers. So that's pretty cool to know.
-Bryan
I could do a search on the database to figure the average sensitivity, however right now I'm having difficulty connecting. I'll try again later.
If the average is down about 80% then that means we at about the level where our smallest CM tracks are at the detection threshold for our set of volunteers. So that's pretty cool to know.
-Bryan
"I am made from the dust of the stars, and the oceans flow in my veins"
- RUSH
- RUSH
The size of the CM tracks may not be the main reason that average sensitivity is down. It is not that I don't miss the small tracks sometimes, but my main problem is with CMs where the "tracks" don't really look like anything. The grey blob on one CM is nothing and on another CM it is a track. It might be worth while is see what CMs are missed most frequently to determine if size of the track is a factor or quality of the CM.If the average is down about 80% then that means we at about the level where our smallest CM tracks are at the detection threshold for our set of volunteers. So that's pretty cool to know.
-
- Posts: 190
- Joined: Fri Aug 04, 2006 4:38 am
- Location: Massachusetts, USA
Hi Ergophobic, Dee and others:
You may want to check out "Post your stats" in the Community section of the forum. Only a handful of us have been brave enough (or good enough) to post our new Phase II stats!
I already have 25 errors, compared to the 12 I had in all of Phase I. Stats are at 100/97.9. I don't think I'm ever going to see 98%, while my Phase I stats were both in the 99% range. Particularly troublesome to me are those microscopic specks, which these aging eyes have trouble seeing. Also have trouble with multiples, which they are trying to trick us with! LOL. Most frustrating of all is when I click "no track" and then see the track.
Well, I guess this is what they wanted. So happy to oblige!
You may want to check out "Post your stats" in the Community section of the forum. Only a handful of us have been brave enough (or good enough) to post our new Phase II stats!
I already have 25 errors, compared to the 12 I had in all of Phase I. Stats are at 100/97.9. I don't think I'm ever going to see 98%, while my Phase I stats were both in the 99% range. Particularly troublesome to me are those microscopic specks, which these aging eyes have trouble seeing. Also have trouble with multiples, which they are trying to trick us with! LOL. Most frustrating of all is when I click "no track" and then see the track.
Well, I guess this is what they wanted. So happy to oblige!