CED 6983 Winter A 2022 Community Influence in Complex Social Networks FNA Influence Mapping: Problem Statement PROBLEM STATEMENT Given a society and use of social media about a topic, in a geographic area of interest, how does an organization identify the right avenue of influence against a specific target (individual or group) through social media or traditional messaging? SUMMARY Utilizing 10,000 social media accounts (twitter) and their joint interest connections, users can identify the right avenues of influence to propagate or stop a message from influencing a society through a specific target (individual or community) OUTCOME Utilizing 1,000 twitter accounts FNA’s application of the community influence algorithm finds the most influential accounts that could best push messages to 20 identified targets of influence campaigns from varying distances across the network. Pictured Left - the original network of 1,000 twitter users in an area of interest FNA Influence Mapping: Sending and Stopping Messages The problem of finding influencers can be mapped to finding the minimal set of structural nodes that if removed would break the information diffusion in the network. The influencers are essential for the structural stability of the information flow. Users elements can choose which entities and paths are best to propagate a message; but can also determine how to remove influencers passing competing messages 20 Target (1,000 twitter users, 3870 connections) network with influencers Without influencers isolated communities are created; enabling enhanced target audience analysis and removing competing messages Showing only influencers identifies which network components are the most susceptible to messaging FNA Influence Mapping: Measures of Performance and Effectiveness The FNA Influence Maximization Solution enables users to identify source of message and target paths, but also predict where information collection should take place to identify onward message propagation (effectiveness) beyond the target and into a target audience. Users can color-code the target’s audience functional subcomponents and identify the most influential nodes in each clique to collect measures of performance and effectiveness indicators The solution can also predict time horizons for message passage; and identify alternative message routes through rewiring capabilities FNA Influence Mapping: Identify Influencers Across Complex Networks Given 20 specifically designated targets for influence in a social network in a defined geographic area of interest: ● 1,000 twitter users are in an target audience network around the 20 identified influence targets ● The targets can analyzed for best approaches individually, or as a group of targets ● Influencers are ranked by their collective capability to project influence across varying distances in the network; Influencer projections are dynamic depending on the network composition ● Collective Influence represent the “influence strength” of a node, containing information on the connectivity of the influenced nodes at a certain distance from source and for the ability the source to spread it. Users can: ● select the best influence points of presence based on placement / proximity to influence target(s) ● determine influence routes based on speed to target vs overt or clandestine situations ● Define a priority list of influencers that is optimal to control (enhance it or break it) the information flow of the network is provided. Red Diamonds = Influencers Blue Asterisk = Influence Target Gold Links = Shortest Paths Blue Link Width = Information Flow Relevance FNA Influence Mapping: Data Science Explanation The problem of finding influencers can be mapped to finding the minimal set of structural nodes that if removed would break the information diffusion in the network. ● Our concept of influencers is based on detecting these structural nodes. ● The whole frame of interconnections in complex networks hinges on a specific set of structural nodes which if activated cause the spreads of information to the whole network. The problem of finding influencers can be mapped to finding the minimal set of structural nodes that if removed would break the information diffusion in the network. ● The influence maximization algo FNA uses is able to detect low degree nodes surrounded by hierarchical coronas of hubs uncovering them through the optimal interplay of all the influencers in the network. The algorithm is based on the measure of the Collective Influence of each nodes. Collective Influence represent the “influence strength” of a node, containing information on the connectivity of the influenced nodes at a certain distance from source and for the ability the source to spread it. ● The algorithm is able to run in linear time and its suitable to find structural nodes (influencers or key player to target to break the network). FNA Influence Mapping: Data & Network Dashboard FNA Influence Mapping: Priority Influencers to Targets Ability to Influence Out to 3 degrees of Separation Ab ili ty to In flu en ce O ut to 9 d eg re es o f S ep ar at io n FNA Influence Mapping: Closest 5 Influencers Per Target Each target is connected to its 5 closest influencers. All the shortest paths between them are highlighted. In some instances, the top 5 influencers create larger components with multiple influence targets Questions?
欢迎咨询51作业君
