Data structure to hold information on the lcs's, paths, and final selected
LCS for a similarity measure. This is used by all the SimilarityMetrics
called for a pair of concepts - we load the lcs info only once for each
concept pair we are comparing.
lcses - set of lcses.
lcsDist - distance between concepts through lcs
lcsPathMap - map of lcs to paths through the lcs between concept pairs. If
this is non-null, then we fill this in. else we ignore this.
corpusLcs, corpusLcsIC - the lcs selected for computing the similarity
(relevant only to Information Content based measures)
intrinsincLcs, intrinsicLcsIC - the lcs selected for computing the similarity
(relevant only to Intrinsic Information Content based measures)