Distributed Trust Algorithms: Difference between revisions
From Konfidi
Jump to navigationJump to search
No edit summary |
No edit summary |
||
| Line 37: | Line 37: | ||
<td>weighted</td> | <td>weighted</td> | ||
<td>avoid cycles, a couple cutoff thresholds</td> | <td>avoid cycles, a couple cutoff thresholds</td> | ||
</tr> | |||
<tr> | |||
<td>[http://www.trustlet.org/wiki/Appleseed Appleseed]</td> | |||
<td></td> | |||
<td></td> | |||
<td></td> | |||
<td></td> | |||
<td></td> | |||
</tr> | </tr> | ||
Latest revision as of 18:30, 23 June 2008
Between people
| Name | Based on psych. research? | Data model | Concatenation of x -> y | Multiple paths | notes |
|---|---|---|---|---|---|
| Konfidi Multiplicative2 | N | [0,1] | (x*y)½ | maximum | |
| TidalTrust | N | 1-10 | |||
| MoleTrust | N | [0,1] | (x*y) | weighted | avoid cycles, a couple cutoff thresholds |
| Appleseed | |||||
| Advogato | Uses Network Flow theory; has central root nodes | ||||
| Patricia Victor's | Y, starting to | bilattice | |||
| OpenPGP | N | untrusted, marginal, full, ultimate |
Between agents
Reputation is generally part of the algorithm
| Valuation of Trust in Open Networks Beth, Borcherding, Klein 1994 |
N | [0,1) | 1-(1-y)x | average | trust values intertwined with # of "good" or "bad" experiences with an entity. Pay attention to "recommendation trust" |
| EigenTrust |
