Hacia una arquitectura flexible para sistemas de predicción de palabraspropuesta de diseño y evaluación

Resumo

Word prediction is a process that tries to guess the word a user is writing, at the same time he/she is doing it. It is mainly used to decrease the effort needed to write a text in applications devoted to people with disabilities. In this paper, we describe a proposal for the architecture of advanced word prediction systems. The proposed architecture is modular and flexible, with common interfaces between the modules to allow the use of different prediction algorithms or even the prediction in a different language. A word prediction system has been implemented according to this architecture. It consists of a 160000 entries lexicon for Spanish (that cannot be modified by the user), the possibility to create and store personal lexicons, prediction methods based on words and POS (parts of speech) probabilistic grammars and a formal grammar. The system can predict more than 94% of the words and reduce the keystrokes needed to write a text more than 50%, depending on the configuration of the prediction. The flexibility of the architecture allowed us to include the word prediction system in several applications for different languages