IMine: a customizable framework for influence mining in complex networks (1.0.0)
IMine is a flexible framework which can be adopt multiple criteria for convergence to solve Influence Minig problems. It can use any diffusion model, as well as resilience to compute the influence of a set of nodes base on the use case. The code is written and tested on ‘R’ v3.5
Installation In order to install this package, you first need to install devtools package: install.packages(“devtools”) Then install this package from GitHub: devtools::install_github(“seekme94/influence.mining”) Package tutorial All influence mining functions are provided in influence_maximization file. The primary interfacing function, influence is a wrapper function, which calls other functions. For most of the users, this should be enough for applications. Example: influence (graph, budget, prob, steps, optimal_solution, test_method, heuristic, centrality_method, parallel, logging) graph is the igraph object budget number of influential nodes to be fetched. Default value is 1 prob the probability at which a node influences its neighbours steps (currently not implemented) is the time steps for which, the diffusion process should run. Provide NULL for exhaustive run. Default value is 1. test_method specifies the method to measure influence. Value MUST be “RESILIENCE”, “INFLUENCE_IC” or “INFLUENCE_LT” heuristic specifies the heuristic method used for influence calculation. Required only when optimal_solution is FALSE centrality_method (optional) is the centrality algorithm to use when heuristic is “CENTRALITY” or “ADAPTIVE_CENTRALITY”. Value must be “DEGREE”, “BETWEENNESS”, “CLOSENESS” or “EIGENVECTOR” parallel when true, executes the funtion using multiple CPU cores. Default value is TRUE optimal solution should be TRUE if influential nodes are to be derived using optimal algorithm. Caution!
This is a companion discussion topic for the original entry at https://www.comses.net/codebases/f4dfc0d3-1634-4ed3-9fdc-c8876ce0832d/releases/1.0.0