Learning circles in social networks.
| dc.contributor.advisor | Vaughn, Randal L. | |
| dc.creator | Ghosh, Debopriya. 1989- | |
| dc.date.accessioned | 2015-05-22T16:16:37Z | |
| dc.date.accessioned | 2017-04-07T19:35:21Z | |
| dc.date.available | 2015-05-22T16:16:37Z | |
| dc.date.available | 2017-04-07T19:35:21Z | |
| dc.date.created | 2015-05 | |
| dc.date.issued | 2015-03-27 | |
| dc.date.submitted | May 2015 | |
| dc.date.updated | 2015-05-22T16:16:38Z | |
| dc.description.abstract | Social networks are ubiquitous. One of the main organizing principles in these real world networks is that of network communities, where sets of nodes organize into densely inked clusters. Identifying such close-knit clusters is crucial for the understanding of the structure as well as the function of these real world networks. We implement an efficient variation of Kernel Spectral Clustering to infer the community affiliation by taking a well represented subgraph of the parent network along with a new notion of cluster mining on feature space to harness the vast amount of rich information stored in users' profile. The proposed method is memory and computationally more efficient than prevalent state-of-art methods. We empirically evaluate our approach against several real world datasets like Facebook, Twitter and Google+ and demonstrate its effectiveness in detecting community affiliations in sparse networks. | |
| dc.format.mimetype | application/pdf | |
| dc.identifier.uri | http://hdl.handle.net/2104/9330 | |
| dc.language.iso | en | |
| dc.rights.accessrights | Worldwide access. | |
| dc.subject | Egonets | |
| dc.subject | Alters | |
| dc.subject | Social networks | |
| dc.subject | Social networking | |
| dc.subject | Kernel Spectral Clustering | |
| dc.title | Learning circles in social networks. | |
| dc.type | Thesis | |
| dc.type.material | text |