Methods of analyzing key entities in a social network

1. In a non-transitory computer-readable medium, a method of generating a subgraph of key entities in a network, segmenting the subgraph of key entities of a network into subgroups based on network centrality metrics, and identifying entities which span the subgroups, the method comprising:

a. generating the subgraph of key entities, comprising:

i. determining, by executing at least one computer program on one or more processors to determine at least two independent types of network centrality metrics for a plurality of entities in a giant component of the network;

ii. selecting starting cut-off criteria for each network centrality metric;

iii. assigning a plurality of rank-based scores to each entity, where each rank based score corresponds to each network centrality metric value of each entity;

iv. selecting some of the entities to form a tentative subgraph of key entities, wherein a subgraph of key entities is defined as a social network graph of a group of entities within the giant component of the network, where the selection of key entities is based on comparisons of the plurality of rank-based scores to the plurality of cut-off criteria, the selection process comprising:

1. identifying the entities whose centrality metric score for a network centrality metric exceeds the cut-off criteria;

2. merging the identified entities into the tentative subgraph; and

3. iteratively repeating steps 1 and 2 of the selection process for each network centrality metric;

v. calculating reach of the entities in the tentative subgraph of key entities and comparing the reach to a reach range, wherein the reach is defined a[s] a fraction of entities in the giant component which are connected to the entities in the tentative subgraph by first-degree links, second-degree links or a combination thereof; wherein the first-degree and second-degree links are based upon co-authorships, co-citations, cross-institutional relationships, or a combination thereof, and where the reach range is defined by a minimum and a maximum reach threshold;

vi. if the reach falls within the reach range, defined as falling above the minimum reach threshold and below the maximum reach threshold, assigning the entities in the tentative sub graph of key entities to the subgraph of key entities, wherein the subgraph of key entities represents a core group of key opinion leaders within a social network;

vii. if the reach does not fall within the reach range, independently adjusting one or more of the cut-off criteria for each network centrality metric; and

viii. iteratively repeating steps iii-vii of the generating the subgraph of key entities until the subgraph of key entities with a reach within the reach range is formed;

b. segmenting the subgraph into subgroups, comprising:

i. calculating, by executing at least one computer program on one or more processors to calculate at least two independent network centrality metric values for each entity in the sub graph of key entities, wherein the network centrality metrics quantitatively indicate independent characteristics of an entity in the network;

ii. assigning a plurality of rank-based scores to each entity, where each rank based score corresponds to each network centrality metric value of each entity;

iii. defining a plurality of rank-based score ranges for each of the network centrality metrics;

iv. defining a plurality of subgroups corresponding to combinations of the rank based score ranges;

v. assigning each entity of the sub graph of key entities to one or more subgroups, comprising:

1. for each rank based score range, determining the score range that the entity's rank-based score falls into;

2. for each subgroup, determining if the entity's rank-based scores fall within the score ranges corresponding to the subgroup;

3. assigning the entity to the corresponding subgroup; and

4. if the entity falls within more than one subgroup, optionally removing the entity from one or more of the subgroups; and

vi. repeating steps iii-v of the segmenting the subgraph into subgroups until a predetermined number of entities is assigned to each subgroup; and

c. identifying a plurality of entities which span the subgroups, comprising:

i. for each entity in each subgroup, determining the shortest path through the network to each entity in each of the other subgroups, resulting in an array of connection paths for each entity;

ii. counting the number of first and second degree connections to members of other subgroups each entity has, wherein a first degree connection is a single-link relationship between an entity and another entity in a different subgroup, wherein a second degree connection is a two-link relationship between an entity and another entity in a different subgroup, resulting in a bridging connection count for each entity; and

iii. selecting a plurality of entities with the highest bridging connection counts.

2. In a non-transitory computer-readable medium, a method of generating a subgraph of key entities in a network, and centers of significance within the network, the method comprising:

a. generating the subgraph of key entities, comprising:

i. determining, by executing at least one computer program on one or more processors to determine at least two independent types of network centrality metrics for a plurality of entities in a giant component of the network;

ii. selecting starting cut-off criteria for each network centrality metric;

iii. assigning a plurality of rank-based scores to each entity, where each rank based score corresponds to each network centrality metric value of each entity;

iv. selecting some of the entities to form a tentative subgraph of key entities, wherein a sub graph of key entities is defined as a social network graph of a group of entities within the giant component of the network, where the selection of key entities is based on comparisons of the plurality of rank-based scores to the plurality of cut-off criteria, the selection process comprising:

1. identifying the entities whose centrality metric score for a network centrality metric exceeds the cut-off criteria;

2. merging the identified entities into the tentative sub graph; and

3. iteratively repeating steps 1 and 2 of the selection process for each network centrality metric;

v. calculating reach of the entities in the tentative sub graph of key entities and comparing the reach to a reach range, wherein the reach is defined as a fraction of entities in the giant component which are connected to the entities in the tentative subgraph by first-degree links, second-degree links or a combination thereof, wherein the first-degree and second-degree links are based upon co-authorships, co-citations, cross-institutional relationships, or a combination thereof, and where the reach range is defined by a minimum and a maximum reach threshold;

vi. if the reach falls within the reach range, defined as falling above the minimum reach threshold and below the maximum reach threshold, assigning the entities in the tentative sub graph of key entities to the subgraph of key entities, wherein the subgraph of key entities represents a core group of key opinion leaders within a social network;

vii. if the reach does not fall within the reach range, independently adjusting one or more of the cut-off criteria for each network centrality metric; and

viii. iteratively repeating steps iii-vii of the generating the subgraph of key entities until the subgraph of key entities with a reach within the reach range is formed;

b. identifying centers of significance within a network, wherein a center of significance is an organization with a high degree of influence, comprising:

i. for each entity in the network, identifying a list of organizations to which the entity belongs;

ii. for each entity in the subgraph of key entities, identifying a list of organizations to which the entity belongs;

iii. counting the number of entities with a membership in each organization;

iv. counting the number of key entities with a membership in each organization;

v. for each organization, counting the number of connections between entities with a membership in the organization;

vi. for each organization, counting the number of connections between members and non-members of the organization;

vii. ranking the organizations according to a formula based on the number of entities and key entities belonging to the organization, and the number of connections within and outside the organization; and

viii. selecting a plurality of the highest ranking organizations; and

c. identifying entities which are important within each center of significance, comprising:

i. for each entity with a membership in each center of significance:

1. counting the number of connections to other entities with a membership in the center of significance; and

2. counting the number of connections to entities who are not members of the center of significance;

ii. ranking the entities according to a formula based on the number of connections within the center of significance and outside the center of significance; and

iii. selecting a plurality of entities with the highest rank.