More than Words: Using Token Context to Improve Canonicalization of Historical German
Bryan Jurish · LDV-Forum/Journal for language technology and computational linguistics · 2010
More Than Words set of candidate hypotheses [w] for a given input word w is non-empty, 2 and it provides a baseline with respect to which the relative utility of more sophisticated conflators can be evaluated.As an example, consider the historical form Abſt e ande, a variant of the contemporary cognate Abstände ("distances").The conflation set [Abſt e ande] id = {Abſt e ande} is nonempty, but does not contain the desired contemporary cognate (Abstände ∈ [Abſt e ande] id ), so Equation (20) from section 4.2 dictates that no instances of the historical variant Abſt e ande will be retrieved via string identity for a query of the contemporary form Abstände.