An Adaptive Word Prediction Interface

Word prediction is a technique often used by individuals with disabilities as a means of increasing the rate with which they can enter text into a computer by reducing the number of keystrokes entered per word. A word prediction system typically operates by presenting a list of "best guesses" for the word the user is currently entering. As the user continues to enter letters, the system updates the list of word predictions to conform to the user’s input. When the word the user is entering is displayed on the list, the user can select the word with one keystroke (often, one of the number keys on the keyboard), and the system will then complete the word for the user. Table 1 lists the parameters defining the word prediction interface that can have an effect on a user’s text entry rate. Koester and Levine [9] demonstrated that while word prediction often reduces the number of keystrokes required to enter text, it does not necessarily follow that it will also reduce the time required to enter that text. The success of word prediction as a text entry rate enhancement technique depends on a variety of factors ineluding the settings of the system parameters described in Table 1, the strategy employed by the user, and the success that the word prediction system has in predicting what words the user is entering. Koester and Levine [10] developed several models that accurately predicted text entry rate enhancement based on these factors, which we are using as the basis for an adaptive word prediction system that automatically modifies its parameters to maximize a user’s text entry rate.