If, during a search for duplicates, a node x, finds another node to which it is more similar than any of it currents neighbours, it should add a connection to the newly found node and remove the connection to the node to which it is least similar (has highest weight). This will make the graph perform faster overtime (less edges to walk to find more similar issues) and increases the chance that other similar nodes (issues) are found through the newly connected nodes.
If, during a search for duplicates, a node x, finds another node to which it is more similar than any of it currents neighbours, it should add a connection to the newly found node and remove the connection to the node to which it is least similar (has highest weight). This will make the graph perform faster overtime (less edges to walk to find more similar issues) and increases the chance that other similar nodes (issues) are found through the newly connected nodes.