Method and device for calculating trust values on purchases

1. A device for calculating trust values of n nodes (u i , i=1 to n), said device comprising:

a processor configured to execute:

a first unit for relating nodes with one another with arrows when the respective nodes have corresponding relations with one another;

a second unit for obtaining P ij (i, j=1 to n) as a weight for an arrow directed from a node u i to a node u i on the basis of the corresponding relation between nodes u i and u i ;

a third unit for determining P′=cP+(1−c)E, where c is a constant, E is a matrix in

which each element is greater than 0 and the sum of each row is 1, P is a matrix in

which an element of row i and column j is represented by P ij ; and

a fourth unit for calculating a trust value TV(u i ) (i=1 to n) of each node u i as an eigen vector for an eigen value 1 of P′ T based on the following equation:

TV

⁡

(

u

i

)

=

∑

u

j

∈

BN

⁡

(

u

i

)

⁢

P

ij

′

⁢

TV

⁡

(

u

j

)

where,

u j εBN(u i ) represents a set of all nodes from which arrows directed at node u i start; and

an output device for outputting information about a node or nodes with high trust values to a user to assist the user in identifying the trustworthy node or nodes among all said n nodes.

2. The device according to claim 1 , wherein

the first unit is arranged for relating nodes u i and u j with zero or more arrows on the basis of the corresponding relationship between nodes u i and u j ; and

for each pair of nodes u i and u j , the second unit is arranged for

assigning initial weight values to the arrows;

calculating a first sum of the initial weight values of all the arrows directed from node u i to node u j ;

calculating a second sum of the initial weight values of all the arrows having node u i as a starting point; and

dividing the first sum by the second sum to obtain P ij .

3. The device according to claim 2 , wherein the third unit is arranged for:

defining a vector v in which elements represent degrees of trust for the respective nodes and the sum of all elements is equal to 1, and

calculating said matrix E=e·v T , and P′=cP+(1−c)E, where c is a constant in 0≦c≦1 and e is a vector in which all elements are equal to 1.

4. The device according to claim 1 , wherein the third unit is arranged for:

defining a vector v in which elements represent degrees of trust for the respective nodes and the sum of all elements is equal to 1, and

calculating said matrix E=e·v T , and P′=cP+(1−c)E, where c is a constant satisfying 0≦c≦1 and e is a vector in which all elements are equal to 1.

5. A computer-readable storage medium or storage device storing therein a program for causing, when executed by a processing arrangement, the processing arrangement to calculate trust values of n nodes (u i , i=1 to n) by executing a process comprising:

relating nodes with one another with arrows when the respective nodes have corresponding relations with one another;

obtaining P ij (i, j=1 to n) as a weight for an arrow directed from a node u i to a node u j on the basis of the corresponding relation between nodes u i and u j ;

determining P′=cP+(1−c)E, where c is a constant, E is a matrix in which each element is greater than 0 and the sum of each row is 1, P is a matrix in which an element of row i and column j is represented by P ij ;

calculating a trust value TV(u i ) (i=1 to n) of each node u i as an eigen vector for an eigen value 1 of p′ T based on the following equation:

TV

⁡

(

u

i

)

=

∑

u

j

∈

BN

⁡

(

u

i

)

⁢

P

ji

′

⁢

TV

⁡

(

u

j

)

where,

u j ε BN(u i )

represents a set of all nodes from which arrows directed at node u i start; and

outputting, via an output device, information about a node or nodes with high trust values to a user to assist the user in identifying the trustworthy node or nodes among all said n nodes.

6. The medium or device according to claim 5 , wherein:

said relating comprises relating nodes u i and u j with zero or more arrows on the basis of the corresponding relationship between nodes u i and u j ; and

said obtaining comprises, for each pair of nodes u i and u j ,

assigning initial weight values to the arrows;

calculating a first sum of the initial weight values of all the arrows directed from node u i to node u j ;

calculating a second sum of the initial weight values of all the arrows having node u i as a starting point; and

dividing the first sum by the second sum to obtain P ij .

7. The medium or device according to claim 6 , wherein said determining comprises:

defining a vector v in which elements represent degrees of trust for the respective nodes and the sum of all elements is equal to 1, and

calculating said matrix E=e·v T , and

P′=cP+(1−c)E, where c is a constant satisfying 0≦c≦1 and e is a vector in which all elements are equal to 1.

8. The medium or device according to claim 5 , wherein said determining comprises:

defining a vector v in which elements represent degrees of trust for the respective nodes and the sum of all elements is equal to 1, and

calculating said matrix E=e·v T , and

P′=cP+(1−c)E, where c is a constant satisfying 0≦c≦1 and e is a vector in which all elements are equal to 1.

9. A method of calculating trust values of n nodes (u i , i=1 to n), said method comprising:

relating nodes with one another with arrows when the respective nodes have corresponding relations with one another;

obtaining P ij (i, j=1 to n) as a weight for an arrow directed from a node u i to a node u i on the basis of the corresponding relation between nodes u i and u i ;

determining P′=cP+(1−c)E, where c is a constant, E is a matrix in which each element is greater than 0 and the sum of each row is 1, P is a matrix in which an element of row i and column j is represented by P ij ; and

calculating, by a processor, a trust value TV(u i ) (i=1 to n) of each node u i as an eigen vector for an eigen value 1 of P′ T based on the following equation:

TV

⁡

(

u

i

)

=

∑

u

j

∈

BN

⁡

(

u

i

)

⁢

P

ij

′

⁢

TV

⁡

(

u

j

)

where,

u j εBN(u i ) represents a set of all nodes from which arrows directed at node u i start; and

outputting, via an output device, information about a node or nodes with high trust values to a user to assist the user in identifying the trustworthy node or nodes among all said n nodes.

10. The method according to claims 9 , wherein:

said relating comprises relating nodes u i and u j with zero or more arrows on the basis of the corresponding relationship between nodes u i and u j ; and

said obtaining comprises, for each pair of nodes u i and u j ,

assigning initial weight values to the arrows;

calculating a first sum of the initial weight values of all the arrows directed from node u i to node u j ;

calculating a second sum of the initial weight values of all the arrows having node u i as a starting point; and

dividing the first sum by the second sum to obtain P ij .

11. The method according to claim 10 , wherein said determining comprises:

defining a vector v in which elements represent degrees of trust for the respective nodes and the sum of all elements is equal to 1, and

calculating said matrix E=e·v T , and P′=cP+(1−c)E, where c is a constant satisfying 0≦c≦1 and e is a vector in which all elements are equal to 1.

12. The method according to claim 9 , wherein said determining comprises:

defining a vector v in which elements represent degrees of trust for the respective nodes and the sum of all elements is equal to 1, and

calculating said matrix E=e·v T , and P′=cP+(1−c)E, where c is a constant satisfying 0≦c≦1 and e is a vector in which all elements are equal to 1.