Loading...
Please wait, while we are loading the content...
Similar Documents
Learning Instance-Level Constraints in Folksonomies for Semi-supervised Clustering using CHR
| Content Provider | Semantic Scholar |
|---|---|
| Author | Shalaby, Maged Abdennadher, Slim Sharaf, Nada Fakhry, Ghada |
| Copyright Year | 2015 |
| Abstract | In our modern days, huge amount of information is available to be searched and browsed by simple users. However, due to the sheer volume of information available, the task of browsing this information becomes increasingly difficult. Clustering algorithms have emerged as an automated tool to organize information for easier browsing. In this paper, we employ constraint reasoning to reason about the different pairwise constraints present, such as must-link and cannot-link constraints. These constraints are used to aid in semi-supervised clustering algorithms, and they are used to guide the algorithm whether a pair of items should or should not be placed in the same cluster. This implementation was done using CHR, a declarative rule-based language. We prove via experimental evaluation that inferring instance-level constraints in the domain of folksonomies yields an improvement in clustering performance. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://met.guc.edu.eg/Repository/Faculty/Publications/531/WLP.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
