Approximate subgraph matching
Apr 22, 2011 3825 views 
en-de
en-es
en-fr
en-pt
en-sl
en
en-zh
0.25
0.5
0.75
1.25
1.5
1.75
2
The paper presents an approach to detection of topic variations based on approximate graph matching. Text items are represented as semantic graphs and approximately matched based on a taxonomy of node and edge labels. Best-matching subgraphs are used as a template against which to align and compare the articles. The proposed approach is applied on news stories using WordNet as the predefined taxonomy. Illustrative experiments on real-world data show that the approach is promising.