In this thesis, a treemap-based interactive clustering algorithm is implemented and evaluated. The treemap's rectangles present scatterplots of dimensionally reduced renderings of clusters. Based on the visual representation, a user makes decisions on splitting clusters with a user-selectable clustering algorithm like k-means or hierarchical clustering. This thesis shows that this implementation is practical for finding hierarchical clusters in the MNIST database. The thesis also demonstrates the possibility of finding sub-clusters within the clusters of the individual digits.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:kau-88044 |
| Date | January 2022 |
| Creators | Lindström-Åberg, Andreas |
| Publisher | Karlstads universitet, Institutionen för matematik och datavetenskap (from 2013), Karlstads universitet, Avdelningen för datavetenskap |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0017 seconds