Abstract: Community detection in graphs identifies groups and is an essential component of graph theory. The clique percolation method (CPM) has been widely used in graph analysis, but there are computation issues when graphs are large. In this study, we use a Venture Capital dataset from 50 years and show the limitations of the k-clique algorithms. Alternatively, we conducted a complete subgraph search for community detection. The computation time and performance of our complete subgraph search are significantly better than the k-clique algorithm.
Keywords—network analysis, community detection, k-cliques
[1]. Alexy, O. T., Block , J. H., Sandner, P., and Ter Wal, A. L. J. 2012. "Social Capital of Venture Capitalists and Start-up Funding," Small Business Economics (39), pp. 835–851.
[2]. Baudin, A. 2021. "Clique Percolation Method: Memory Efficient Almost Exact Communities."
[3]. Hegde, D., and Tumlinson, J. 2014. "Does Social Proximity Enhance Business Partnerships? Theory and Evidence from Ethnicity's Role in U.S. Venture Capital," Management Science (60:9), pp. 2355-2380.
[4]. Mote, J. 2013. "Syndication, Networks and the Growth of Venture Capital in Philadelphia, 1980–99," Industry and Innovation (18:1), pp. 131–150.
[5]. Noyes, E., Brush, C., Hatten, K., and Smith-Doerr, L. 2014. "Firm Network Position and Corporate Venture Capital Investment," Journal of Small Business Management (52:4), pp. 713-731.