Collaborative networking with optimized inter-domain information quality assessment

1. A method for implementing inter-domain information quality assessment, comprising:

calculating, at a computer, relationships among domains in a knowledge base by identifying features that respectively define each of the domains and identifying domains having correlating features;

assigning, at the computer, values to pairs of the domains that reflect a closeness of the relationships based upon the correlating features;

evaluating, at the computer, a reputation of an entity in one domain based on past performance of the entity with respect to another domain where the other domain is determined, via the assigned values, to be related to the domain, the past performance of the entity indicating capabilities of the entity to render accurate predictions about events in a particular domain;

calculating, at the computer, a value representing the reputation of the entity with respect to the domain based upon the evaluating; and

assigning the value representing the reputation to the entity.

2. The method of claim 1 , wherein the domain represents a subject in which there is no a priori knowledge of predictions rendered by the entity.

3. The method of claim 1 , further comprising evaluating, at the computer, the reputation of the entity in the domain based on past performance of the entity with respect to the domain.

4. The method of claim 1 , further comprising:

receiving, at the computer, a value relating to a predicted outcome of an event from the entity and at least one other entity, the event associated with a domain;

receiving, at the computer, a value representing the reputation for the entity and a value representing a reputation for the at least one other entity with respect to the domain;

receiving, at the computer, inter-domain distance metrics specifying the values assigned to the pairs of the domains reflecting a closeness of the relationships;

calculating, at the computer, a weighted average of the values received relating to the predicted outcome for the event, where the weighted average is based upon reputation scores associated with the entity and the at least one other entity with respect to the domain, the inter-domain distance metrics, and reputation scores of the entity and the at least one other entity for others of the domains determined to be related to the domain via the inter-domain distance metrics; and

generating, at the computer, a collective prediction value for the event that reflects a collective prediction for the event, the collective prediction value generated from calculating the weighted average of the values.

5. The method of claim 4 , further comprising:

comparing, at the computer, the value relating to the predicted outcome for each of the entity and the at least one other entity with a value specifying an actual outcome for the event with respect to the domain;

determining, at the computer, a prediction error value for each of the entity and the at least one other entity with respect to the domain in response to the comparing, the prediction error value reflecting a difference between the value of the predicted outcome and the value of the actual outcome; and

using the prediction error value for the domain, assigning or updating, at the computer, the reputation scores for the entity and the at least one other entity with respect to the domain and respective related domains.

6. The method of claim 4 , further comprising:

updating, at the computer, the inter-domain distance metrics based upon the predicted outcome of the event.

7. The method of claim 4 , wherein the weighted average is determined by a function:

p

d

⁡

(

t

)

=

∑

i

⁢

w

i

,

t

-

1

d

⁢

x

i

d

⁡

(

t

)

∑

i

⁢

w

i

,

t

-

1

d

;

wherein t represents a decision time during which predictions are collected for the event, d represents a domain, w represents a reputation score, i represents an index for the entity, and w i,t-1 represents a reputation score of an entity i at decision time t.

8. The method of claim 1 , wherein the value representing the reputation is derived by a function:

w

i

,

t

j

=

∑

d

=

1

,

…

⁢

,

D

;

d

≠

j

⁢

α

jd

⁢

∂

Φ

∂

R

i

,

t

d

;

wherein t represents a decision time during which predictions are collected for the event, y represents an outcome of the event, d represents a domain, D represents the number of domains, w represents a reputation score, i represents an index for the entity, and R represents a measure of accuracy of a prediction.

9. A computer program product for use by an inter-domain information quality assessment program for implementing inter-domain information quality assessment optimization, the computer program product comprising:

a non-transitory storage medium readable by a processing circuit and storing instructions for execution by the processing circuit for performing a method comprising:

calculating relationships among domains in a knowledge base by identifying features that respectively define each of the domains and identifying domains having correlating features;

assigning values to pairs of the domains that reflect a closeness of the relationships based upon the correlating features;

evaluating a reputation of an entity in one domain based on past performance of the entity with respect to another domain where the other domain is determined, via the assigned values, to be related to the domain, the past performance of the entity indicating capabilities of the entity to render accurate predictions about events in a particular domain;

calculating a value representing the reputation of the entity with respect to the domain based upon the evaluating; and

assigning the value representing the reputation to the entity.

10. The computer program product of claim 9 , wherein the domain represents a subject in which there is there is no a priori knowledge of predictions rendered by the entity.

11. The computer program product of claim 9 , wherein the method further comprises evaluating the reputation of the entity in the domain based on past performance of the entity with respect to the domain.

12. The computer program product of claim 9 , wherein the method further comprises:

receiving a value relating to a predicted outcome of an event from the entity and at least one other entity, the event associated with a domain;

receiving a value representing the reputation for the entity and a value representing a reputation for the at least one other entity with respect to the domain;

receiving inter-domain distance metrics specifying the values assigned to the pairs of the domains reflecting a closeness of the relationships;

calculating a weighted average of the values received relating to the predicted outcome for the event, where the weighted average is based upon reputation scores associated with the entity and the at least one other entity with respect to the domain, the inter-domain distance metrics, and reputation scores of the entity and the at least one other entity for others of the domains determined to be related to the domain via the inter-domain distance metrics; and

generating a collective prediction value for the event that reflects a collective prediction for the event, the collective prediction value generated from calculating the weighted average of the values.

13. The computer program product of claim 12 , wherein the method further comprises:

comparing the value relating to the predicted outcome for each of the entity and the at least one other entity with a value specifying an actual outcome for the event with respect to the domain;

determining a prediction error value for each of the entity and the at least one other entity with respect to the domain in response to the comparing, the prediction error value reflecting a difference between the value of the predicted outcome and the value of the actual outcome; and

using the prediction error value for the domain, assigning or updating the reputation scores for the entity and the at least one other entity with respect to the domain and respective related domains.

14. The computer program product of claim 12 , wherein the method further comprises:

updating the inter-domain distance metrics based upon the predicted outcome of the event.

15. The computer program product of claim 12 , wherein the weighted average is determined by a function:

p

d

⁡

(

t

)

=

∑

i

⁢

w

i

,

t

-

1

d

⁢

x

i

d

⁡

(

t

)

∑

i

⁢

w

i

,

t

-

1

d

;

wherein t represents a decision time during which predictions are collected for the event, d represents a domain, w represents a reputation score, i represents an index for the entity, and w i,t-1 represents a reputation score of an entity i at decision time t.

16. The computer program product of claim 9 , wherein the value representing the reputation is derived by a function:

w

i

,

t

j

=

∑

d

=

1

,

…

⁢

,

D

;

d

≠

j

⁢

α

jd

⁢

∂

Φ

∂

R

i

,

t

d

;

wherein t represents a decision time during which predictions are collected for the event, y represents an outcome of the event, d represents a domain, D represents the number of domains, w represents a reputation score, i represents an index for the entity, and R represents a measure of accuracy of a prediction.

17. A system for implementing inter-domain information quality assessment optimization, the system comprising:

a computer; and

an inter-domain information quality assessment application executable by the computer, the inter-domain information quality assessment application configured to perform a method comprising:

calculating relationships among domains in a knowledge base by identifying features that respectively define each of the domains and identifying domains having correlating features;

assigning values to pairs of the domains that reflect a closeness of the relationships based upon the correlating features;

evaluating a reputation of an entity in one domain based on past performance of the entity with respect to another domain where the other domain is determined, via the assigned values, to be related to the domain, the past performance of the entity indicating capabilities of the entity to render accurate predictions about events in a particular domain;

calculating a value representing the reputation of the entity with respect to the domain based upon the evaluating; and

assigning the value representing the reputation to the entity.

18. The system of claim 17 , wherein the domain represents a subject in which there is no a priori knowledge of predictions rendered by the entity.

19. The system of claim 17 , wherein the method further comprises evaluating the reputation of the entity in the domain based on past performance of the entity with respect to the domain.

20. The system of claim 17 , wherein the method further comprises:

receiving a value relating to a predicted outcome of an event from the entity and at least one other entity, the event associated with a domain;

receiving a value representing the reputation for the entity and a value representing a reputation for the at least one other entity with respect to the domain;

receiving inter-domain distance metrics specifying the values assigned to the pairs of the domains reflecting a closeness of the relationships;

calculating a weighted average of the values received relating to the predicted outcome for the event, where the weighted average is based upon reputation scores associated with the entity and the at least one other entity with respect to the domain, the inter-domain distance metrics, and reputation scores of the entity and the at least one other entity for others of the domains determined to be related to the domain via the inter-domain distance metrics; and

generating a collective prediction value for the event that reflects a collective prediction for the event, the collective prediction value generated from calculating the weighted average of the values.

21. The system of claim 20 , wherein the method further comprises:

comparing the value relating to the predicted outcome for each of the entity and the at least one other entity with a value specifying an actual outcome for the event with respect to the domain;

determining a prediction error value for each of the entity and the at least one other entity with respect to the domain in response to the comparing, the prediction error value reflecting a difference between the value of the predicted outcome and the value of the actual outcome; and

using the prediction error value for the domain, assigning or updating the reputation scores for the entity and the at least one other entity with respect to the domain and respective related domains.

22. The system of claim 20 , wherein the method further comprises:

updating the inter-domain distance metrics based upon the predicted outcome of the event.

23. The system of claim 20 , wherein the weighted average is determined by a function:

p

d

⁡

(

t

)

=

∑

i

⁢

w

i

,

t

-

1

d

⁢

x

i

d

⁡

(

t

)

∑

i

⁢

w

i

,

t

-

1

d

;

wherein t represents a decision time during which predictions are collected for the event, d represents a domain, w represents a reputation score, i represents an index for the entity, and w i,t-1 represents a reputation score of an entity i at decision time t.

24. The system of claim 17 , wherein the value representing the reputation is derived by a function:

w

i

,

t

j

=

∑

d

=

1

,

…

⁢

,

D

;

d

≠

j

⁢

α

jd

⁢

∂

Φ

∂

R

i

,

t

d

;

wherein t represents a decision time during which predictions are collected for the event, y represents an outcome of the event, d represents a domain, D represents the number of domains, w represents a reputation score, i represents an index for the entity, and R represents a measure of accuracy of a prediction.