semantic-similarity.txt

Applications of semantic similarity

-> Grouping similar words into semantic concepts. 
-> In paraphrasing - Rewrite one sentence to another with same meaning
-> Textual entailment - Smaller sentence entails its meaning from another piece of text (derives its meaning)

Resource used - WordNet (For English)


WordNet organizes information as a tree. 

-> Shortest path between two concepts gives semantic similarity (Path similarity)
Closer the two concepts are, more is the semantic similarity. (Inversely proportional to the distance).

-> Lowest Common Subsumer (LCS) - Lin similarity
Find the closest ancestor to both the concepts

Lin similarity(u,v) = 2*log(P(LCS(u,v)))/ (log(P(u)) + log(P(v)))


->Collocations and Distributional similarity -
Two words appearing in similar context are more likely to be semantically related. 

Compute strength of association of 2 words - 
> How frequently they occur? Not similar if they don't occur together often,
> How frequently the individual words occur? 'The' occurs very frequent.
> Pointwise Mutual Information 
	PMI(w,c) = log(P(w,c)/P(w)P(c))