..........................................NEXT PREVIOUS INDEX
L04 Morphology - Using Tries Morphology: Towards Finite State Transducers o Trie data structure l k | | a a | | D p | | -------- a | | | a A D | | | k ------- | | | | ------------ | I i | | | ------- | A e o | | | A | | | | | | A e o M M | M o Abstracting out suffixes k l | | a a | | p D | | a --------- | | | D(#1) a A | | k (#1) I #1: Corresponds to paradigm for 'laDakA' - Suffix trie (forward) #1 | -------- | | e o | M - Alternatively use suffix trie backward Suffix Trie 0 | --------- | | e (#1) M | ------- | | e(#1) 0(#1) #1: Corresponds to paradigm for 'laDakA' o Finite state machines (alternative view of above disjointed trees) (Putting characters on edges rather than nodes.) + / \ l / \ k + + a | | a | | + + D | | p | | + + a | | a | | + + k | | D | | + + \ / 0 \ / 0 +______ e/ \o \ A / \ \ (+) + (+) | |M (+) Legend: - Nodes are shown using '+' - Accepting nodes are shown using '(+)' - Each arc is a directed arc from above to below - Empty string as arch label is shown using '0' But the answer should tell us: - Not just whether the input is a valid word, - But its analysis: + Ex.: 'ladakoM' has root 'ladakA' with number plural, etc. o For this we need finite state transducer...........................................NEXT PREVIOUS INDEX