. . . . . . . . . . . "Community detection plays a pivotal role in social\r\nnetwork analysis by partitioning networks into cohesive groups\r\nof vertices with dense intra-group connections and sparse intergroup connections. In this paper, we utilized a scholarly social\r\nnetwork based on researchers’ topic similarity derived from\r\ntheir publication metadata to identify interdisciplinary research\r\ncommunities. As topics often form a hierarchy, we hypothesize\r\nthat the constructed scholarly network will exhibit hierarchical\r\ncommunity structures. Therefore, we explore the efficacy of two\r\nprominent community detection algorithms, Louvain and Spectral clustering, known for their capacity to detect hierarchical\r\ncommunity structures within networks. While both algorithms\r\ndemonstrate this capability, the original Louvain algorithm is\r\nsusceptible to the resolution limit problem due to its reliance on\r\nthe modularity measure. To address this limitation, we propose\r\nthe nested hierarchical Louvain algorithm, which iteratively\r\npartitions the network based on previously identified subgraphs,\r\nand we find that the bias towards large communities is mitigated.\r\nTo evaluate the hierarchy produced by each of the algorithms,\r\nwe employ the Cophenetic Correlation Coefficient (CPCC), a\r\nmetric commonly used in hierarchical clustering evaluations\r\nbut less frequently utilized in hierarchical community analysis.\r\nWe argue that CPCC can be a useful measure to identify\r\nthe presence of implicit hierarchical community structure in\r\nsocial networks when it is not explicitly available from domain\r\nknowledge while also further mitigating the inherent bias present\r\nin using modularity as a metric. Experimental results, conducted\r\non both synthetic networks and the scholarly social network,\r\ndemonstrate that the nested hierarchical Louvain algorithm, as\r\nwell as Spectral Clustering, successfully identifies more finely\r\nstructured hierarchical communities, offering greater depth in\r\nthe dendrogram compared to the basic Louvain algorithm.\r\nIndex Terms—Social Networks, Hierarchical Community Detection, Clustering, Topic Models" . "Identifying Hierarchical Community Structures in Content-Based Scholarly Social Networks" . . . "john.sheppard@montana.edu" . "2025-03-04" . "2023" . . "Emily Regalado" . "2026-02-04T18:00:04.226Z"^^ . . . . . . . . . "RSA" . "MIIBIjANBgkqhkiG9w0BAQEFAAOCAQ8AMIIBCgKCAQEAxzr6UBGMW6c8tegz0babaledWUEQ0PLDE4tp7Iinbe2DZtAtY5JUptKYuStWDZx+QER4808P8dejNWRnBDzgthYJm/AyNSXflHSJhz2+NC+h7RylOLxbwLEQocmyKKiYxa2gT85m6ajVL2M6TnfG67nnK+K2f7iCGL6wYXRITD1q+7+5SWqBdDXIV921W4IKWaD2GJk+NRBoOqQhbsrk8Tn5XsNd7DMYVHk47oMDGbeBnrOIoRPsbBgAcoCsxxhiB9yN6Lf8EUbnlXVEDzJuZk048L1BDZL+6nkA8btTQGP2ijUFWA7rTrod3LjUDQWLZS95njjl867dtmv/znYkzwIDAQAB" . "fkK51Bxk/kpAi/xzeR+/jziwQtmabO5O7eMYmxjtxqSz+V6XiBCVAKMTrJKIWjHaoG4bdMip97ywI2MBOwJI5V5J04gWqzaOHvZR5l8LeUYF+rMvkQRksam9GpTi5PWj4pxOKCpIvv693qmlz/8Tnon2+gJl7ABgEkNSzqoGLX/53PLo0btDV44o3coQ8567PeXwpyYTPZufJWtDCCa5byTrSPef2TFlCDSfFBotKbD0xWPxKWodss7fLHVvWH1Bh4KXDswYt9Yi34M9jLXX941H0T+S0eweZYR/4BVFn1cnHma03DmEKJiV07ka2wRQU9uKBYRt//j/q79bcvykig==" . . .