Pairwise Semantic Similarity Measure

Context:
- It can range from being a Pairwise (All-Pairs) Semantic Similarity Measure to being a Pairwise (Best-Pairs) Semantic Similarity Measure.
Example(s):
Counter-Example(s):
See: Semantic Similarity Neural Network, Ontology-based Semantic Similarity Measure, Statistical Semantic Similarity, Semantic Word Similarity Measure.

References

(Pesquita et al., 2009 ) ⇒ Catia Pesquita, Daniel Faria, Andre O. Falcao, Phillip Lord, and Francisco M. Couto (2009). "Semantic Similarity in Biomedical Ontologies". In: PLoS Computational Biology 5(7): e1000443.
- QUOTE: Pairwise approaches measure functional similarity between two gene products by combining the semantic similarities between their terms. Each gene product is represented by its set of direct annotations, and semantic similarity is calculated between terms in one set and terms in the other (using one of the approaches described previously for comparing terms). Some approaches consider every pairwise combination of terms from the two sets (all pairs technique), while others consider only the best-matching pair for each term (best pairs technique). A global functional similarity score between the gene products is obtained by combining these pairwise semantic similarities, with the most common combination approaches being the average, the maximum, and the sum.
  (...)

**Table 2:** Summary of pairwise approaches.
^aLei et. al also consider exact matches only.
Measure	Approach	Techniques	Term Comparison
Lord et al. (2003)	All pairs	Average	Resnik/Lin/Jiang
Sevilla et al. (2005)	All pairs	Maximum	Resnik/Lin/Jiang
Riensche et al., (2007, XOA)	All pairs	Maximum	XOA
Azuaje et al. (2005)	Best pairs	Average	Resnik/Lin/Jiang
Couto et al. (2005)	Best pairs	Average	GraSM+(Resnik/Lin/Jiang)
Schlicker et al. (2006, funSim)	Best pairs	Average	simRel
Wang et al. (2007)	Best pairs	Average	Wang
Tao et al. (2007) (ITSS)	Best pairs	Average Min. threshold	Lin
Pozo et al. (2008)	Best pairs	Average	Pozo
Lei et al. (2006)	All pairs Best pairs^a	Average Max, Sum	Depth of LCA

(Cheng et al., 2004) ⇒ Jill Cheng, Melissa Cline, John Martin, David Finkelstein, Tarif Awad, David Kulp, and Michael A. Siani-Rose (2004). "A Knowledge-Based Clustering Algorithm Driven by Gene Ontology". In: Journal of biopharmaceutical statistics, 14(3), 687-700.
- QUOTE: To quantify the relationship between two GO tiodes, we calculate a pair-wise similarity score according to the number of edges two nodes share In their respective paths to the root of the hierarchy. Only the edges that are common to both paths are significant; the edges below their lowest-level common ancestor do not contribute to similarity. Because of multiple inheritance, one node may have multiple parents and may reside in multiple paths, We took a greedy approach, representing the similarity between two nodes as their longest common partial path from the root.