Node relevance determination in an evolving network

A system effective to generate relevancy data for at least one node among a heterogeneous network that changes with time, the system comprising:

• a memory configured to store a database including records related to at least one entity;
• a processor configured to be in communication with the memory, the processor is configured to:
• generate a series of temporal heterogeneous networks based on the database stored in the memory, wherein each temporal heterogeneous network is a state of a heterogeneous network at a time window, wherein the heterogeneous network changes with time, and each temporal heterogeneous network includes a respective set of agent nodes and a respective set of non-agent nodes;
• for each temporal heterogeneous network:
• extract a set of agent nodes;
• generate a relationship network based on the extracted agent nodes, wherein the relationship network includes the extracted agent nodes, and the relationship network corresponds to the temporal heterogeneous network;
• for an agent node among each pair of temporal heterogeneous network and relationship network:
• remove the agent node from the temporal heterogeneous network;
• generate a conditional relationship network based on the remaining agent nodes among the temporal heterogeneous network after the removal of the agent node;
• determine a relevancy score for the removed agent node based on the relationship network and based on the conditional relationship network, wherein the relevancy score corresponds to the time window;
• and
• generate the relevancy data based on the relevancy scores for each agent node among each temporal heterogeneous network, wherein the relevancy data indicates changes of the relevancy scores of each agent node with respect to time.

A computer program product for generating relevancy data of at least one node among a heterogeneous network that changes with time, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processing element of a device to cause the device to:

• generate a series of temporal heterogeneous networks, wherein each temporal heterogeneous network is a state of a heterogeneous network at a time window, wherein the heterogeneous network changes with time, and each temporal heterogeneous network includes a respective set of agent nodes and a respective set of non-agent nodes;
• for each temporal heterogeneous network:
• extract a set of agent nodes;
• generate a relationship network based on the extracted agent nodes, wherein the relationship network includes the extracted agent nodes, and the relationship network corresponds to the temporal heterogeneous network;
• for an agent node among each pair of temporal heterogeneous network and relationship network:
• remove the agent node from the temporal heterogeneous network;
• generate a conditional relationship network based on the remaining agent nodes among the temporal heterogeneous network after the removal of the agent node;
• determine a relevancy score for the removed agent node based on the relationship network and based on the conditional relationship network, wherein the relevancy score corresponds to the time window;
• and
• generate the relevancy data based on the relevancy scores for each agent node among each temporal heterogeneous network, wherein the relevancy data indicates changes of the relevancy scores of each agent node with respect to time.