Supporting Textual Input by Using Multiple Entropy Models

Authors:
Joscha Bach
Affiliations:
Institute of Informatics, Humboldt University of Berlin, Unter den Linden 9, 10199 Berlin, Germany
Venue:
Fundamenta Informaticae
Year:
2001

Citing 2
Cited 0

Lexical Attraction for Text Compression

DCC '99 Proceedings of the Conference on Data Compression
Discovery of linguistic relations using lexical attraction

Discovery of linguistic relations using lexical attraction

Quantified Score

Hi-index	0.00

Visualization

Abstract

The efficiency of textual input of disabled users can be considerably improved by em-ploying prediction methods as they are used in data compression algorithms. In this paper, the application of the PPM algorithm is described along with a technique of utilizing multiple concur-rent prediction models. In a prototypical implementation called PRIS the method has been found to reduce the number of keystrokes necessary to enter textual data by 55 to 62 percent over common text entry. Compared with zero-order prediction as used in most textual input aids that are currently used, PRIS reduces the number of keystrokes between 12 and 24 percent.