Language Models for Text Entry
Kumiko Tanaka-Ishii, University of Tokyo, Tokyo, Japan
For many years, the design of text entry methods has reflected the statistics of language. For example, the Dvorak keyboard, proposed as an alternative to the standard Qwerty keyboard, was designed based on statistics concerning letter usage in the English language. Modeling the language better enables the construction of better entry systems.
In recent years, huge collections of electronic text have appeared and we have more opportunity to acquire language statistics. A body of text that is collected and organized under particular criteria is called a corpus. Gigabytes of newspaper corpora are now available, and international ...