1) Path length Algorithm
When concepts are organized in a hierarchy, it is convenient to measure similarity according to structural measures that find path lengths between concepts. In fact, there has been a variety of such approaches proposed in general English. Rada, et al. (1989) [8], developed a measure based on path lengths between concepts in the WordNet hierarchy Shortest path measure focuses on the closeness of two concepts in the hierarchy.
The shorter path between two words/senses in a thesaurus hierarchy graph, the more similar they are: x Words are quite similar to parents & siblings x Less similar to words far away in the network Pathlen(c1, c2) = number of edges in shortest path x Path-based similarity often involves a log transform x path-length based similarity:
simpath(c1, c2) = −log pathlen(c1, c2)