semantic-similarity/similarity

similarity

Measuring Semantic Similarity in the Taxonomy of Wordnet (sim max b)

Finding graph

Currently, taxonomy is browsed breathly to find shorthest path between concepts by finding common concept, which is one of hypernym, holonym or meronym. All of them are treated as the same ones in criterion of similarity.

Output

For sim_max_b('cat', 'dog') output is: PATHLEN: 4, what is sum of 2+2. Common concepts are both Synset('carnivore.n.01') and Synset('male.n.02'), that are on distance 2 from dog synsets as well as distance 2 from cat synsets

At the distance of 1 from 'cat': [Synset('feline.n.01'), Synset('man.n.01'), ... , Synset('flog.v.01'), Synset('excrete.v.01')]

At the distance of 2 from 'cat': [Synset('carnivore.n.01'), ... , Synset('male.n.02'), Synset('communicator.n.01'), ... , Synset('imaging.n.02'), Synset('beat.v.02'), Synset('exhaust.v.05')]

At the distance of 1 from 'dog': [Synset('canine.n.02'), Synset('domestic_animal.n.01'), ... , Synset('hotdog.n.02'), Synset('ratchet.n.01')]

At the distance of 2 from 'dog': [Synset('carnivore.n.01'), ... , Synset('male.n.02'), Synset('unwelcome_person.n.01'), ... , Synset('mechanical_device.n.01'), Synset('spiral_ratchet_screwdriver.n.01')]