Figure 3
The network diagram displays a clustered citation or keyword network arranged diagonally from upper left to lower right, with connected nodes and labeled directional arrows indicating thematic areas. At the upper left, a dense cluster of nodes includes labels such as “kokina j, 2017, j emerg techno”, “alles mg, 2016, account horiz”, “sutton sg, 2016, int j account”, “warren jd, 2015, account”, and “moll j, 2019, brit account rev”. A thick blue arrow labeled “Using Artificial Intelligence” points leftward toward this upper-left cluster. Moving diagonally downward toward the center, a green cluster includes nodes such as “eining mm, 1997, auditing j pr”, “perols jl, 2017, account rev”, “bao y, 2020, j account res, v5”, “loughran t, 2016, j account re”, “echow pm, 2011, contemp accou”, and “kirkos e, 2007, expert syst ap”. A blue arrow labeled “Using Machine Learning” points rightward toward this central green cluster. Toward the mid-right region, a yellow-toned cluster appears near nodes such as “fanning k, 1998, internati” and “he hb, 2009, ieee t knowl data”. A blue arrow labeled “Neural network” points leftward toward this cluster. Further down toward the lower-right region, a red cluster includes nodes such as “altman ei, 1968, j financ v23”, “chawla nv, 2002, j artific intel”, “bhattacharyya s, 2011, decis s”, and “bahnsen ac, 2016, expert syst”. A blue arrow labeled “Fraud detection” points leftward toward this lower-right cluster. Curved connecting lines run between nodes across all clusters, forming a continuous network that transitions from artificial intelligence–focused studies at the top-left through machine learning and neural network areas toward fraud detection applications at the bottom-right.References co-citation network. Source: The authors