11 Oct 2005
CS 5244 - Computational Document
Analysis
28
Design-based methods
¡Idea:
capture syntactic and semantic flow rather than
token identity (for source code)
¡Replace
variable names with IDs correlated with symbol
table and data type
¡Decompose
each p into
regions of
lsequential
statements
lconditionals
llooping
blocks – recurse on these
¡Calculate
similarity from root node downwards