Evaluación Atomática de un Sistema Híbrido de Predicción de Palabras y Expansiones

Word prediction is one of the most commonly used systems to help to write people with physical and/or linguistic disabilities. In the newest systems, word prediction is complemented with other strategies to improve its performance, such as abbreviation expansion or phrase prediction. In this paper, a hybrid system with prediction of words and expansions is presented. Expansion prediction consists in expanding the abbreviation even before the user finishes writing it. This system allows the user to abbreviate or not a word, and reduces the cognitive load required for its use because it is not necessary to remember a fixed abbreviation for each word. The parameter used to evaluate the efficiency of the system is the percentage of keystrokes saved with respect to writing the text without help, and we include results of the word prediction, the expansion prediction and the combination of both.