Corpus-based stochastic finite-state predictive text entry for reduced keyboards: application to Catalan.
Mikel L. Forcada · RUA, Repositorio Institucional de la Universidad de Alicante (Universidad de Alicante) · 2001
Users of digital mobile \t\t\t\t phones have access to an increasing \t\t\t\t number of text-based services but \t\t\t\t most of them can only use a very reduced \t\t\t\t keyword for text entry. Traditionally, \t\t\t\t this has been solved by means \t\t\t\t of multiple tapping and delays or next-character \t\t\t\t keys (e.g. 66(pause)666 for \t\t\t\t no), which is slow (two taps per letter \t\t\t\t on average), inconvenient, and prone \t\t\t\t to error. Recently, some mobile terminals \t\t\t\t offer dictionary-based predictive \t\t\t\t text-entry schemes which, in almost \t\t\t\t all cases, allow for one-tap-per-letter \t\t\t\t text entry (e.g. 367 for for). This \t\t\t\t paper shows how text corpora can be \t\t\t\t used to build stochastic finite-state automata \t\t\t\t which may in turn be easily \t\t\t\t used to implement predictive text entry \t\t\t\t for medium-sized languages such \t\t\t\t as Catalan, not currently supported \t\t\t\t by major predictive-text-entry companies.