meta user
network profile
| bio | website | |
|---|---|---|
| location | ||
| age | ||
| visits | member for | 1 year, 4 months |
| seen | Dec 18 '12 at 13:13 | |
| stats | profile views | 2 |
|
Jan 25 |
comment |
Aggregate opinions to a single measure @gung, this is really helpful! I'd been thinking about both answers from Michelle and onestop before I saw yours, and I'd already started questioning myself about those things you succinctly pointed out. I will read through the links above, and re-visit my original question and Michelle and onestop's answers. Thanks! |
|
Jan 24 |
comment |
Aggregate opinions to a single measure It is an interesting approach, but a potential problem is that if you have 5 low-skill experts who endorse the same rating versus a high-skill expert who rate differently, then it is possible that the sum of the informativeness scores of the 5 low-skill experts may exceed the high-skill one, and the rating picked by the high-skill expert is not taken into account in any way in the final result. |
|
Jan 24 |
comment |
Aggregate opinions to a single measure thanks for your answer. If I understand correctly, you sum the expert informativeness scores behind each given rating, and pick out the one that has the largest sum, i.e. the one that gathers the highest level of informativeness. |
|
Jan 23 |
comment |
Aggregate opinions to a single measure @Michelle, the example you gave is the idea here: if an expert $a$ only has an informativeness score that is half as large as the one of another expert $b$, then by design, we consider expert $a$ is only half as experienced/skillful in wining tasting as expert $b$, and we would like to reflect this in some way in our aggregation. |
|
Jan 23 |
comment |
Aggregate opinions to a single measure @onestop, as I mentioned above, the problem with the (arithmetic) weighted mean is that, the contribution to the aggregated measure of a low-score high-value rating (e.g. a rating of 5 with an informativeness score 0.05/1) might be equivalent or even higher than a high-score low-value rating (e.g. a rating of 1 with an informativeness score 0.25 or a rating of 1 with an informativeness score 0.2/1). And I think that is unreasonable considering that the latter ratings are 5 or 4 times more informative than the first one, based on the informativeness measure. |
|
Jan 23 |
awarded | Student |
|
Jan 23 |
comment |
Aggregate opinions to a single measure @Michelle, it's (2) mostly: if an expert is regarded more reliable (i.e. with a high informativeness score), then he/she is regarded as more experienced and skillful in judging the wine quality, hence I believe his/her rating would be more accurate. Thanks. |
|
Jan 23 |
awarded | Editor |
|
Jan 23 |
revised |
Aggregate opinions to a single measure deleted 5 characters in body |
|
Jan 23 |
asked | Aggregate opinions to a single measure |
|
Jan 16 |
awarded | Scholar |
|
Jan 16 |
accepted | Different ways of making a set of numbers (all between $0$ and $1$) to sum up to $1$ |
|
Jan 10 |
comment |
Different ways of making a set of numbers (all between $0$ and $1$) to sum up to $1$ @vinux, i saw yours too. thanks for the input. |
|
Jan 10 |
comment |
Different ways of making a set of numbers (all between $0$ and $1$) to sum up to $1$ @Arthur, thanks for the generalization, could you please elaborate with an example for what you mean by "bending" the set $S$? Appreciated. |
|
Jan 10 |
asked | Different ways of making a set of numbers (all between $0$ and $1$) to sum up to $1$ |
