A Self-Reliant finite Automata for Reduplication Detection

Reduplication is a common phenomenon in almost all human languages. It implies the repetition of the smallest linguistic unit partially (e.g. flip flop) or fully (e.g. bye bye). Symbolically it can be written as WiWj where W-i and W(j )are similar or almost similar and W is any linguistic unit or token or sequence of symbols. Identifying such reduplication in a text is an important aspect of various language processing tasks. Several researchers have tried to address the issue and solve using various techniques like the heuristic approach, 1-way / 2-way finite-state transducers, etc. This paper presented a sophisticated generic (Nondeterministic Finite Automata i.e. NFA-based) automaton that is self-reliant generates dynamically and identifies all the reduplicated words from the text.