Alternate Title

Improved text entry for mobile devices

Advisor(s)

Peter Tarasewich

Contributor(s)

Harriet J. Fell, Javed A. Aslam, Carole D. Hafner, Scott I. MacKenzie

Date of Award

2008

Date Accepted

10-2007

Degree Grantor

Northeastern University

Degree Level

Ph.D.

Degree Name

Doctor of Philosophy

Department or Academic Unit

College of Computer and Information Science.

Keywords

Computer Science, Information Science, Mobile Devices

Subject Categories

Cellular telephones, Keyboards (Electronics), Human-computer interaction

Disciplines

Computer Sciences

Abstract

Despite the ever-increasing popularity of mobile devices, text entry on such devices is becoming more of a challenge. Problems primarily lie with shrinking device sizes, which can greatly limit available display space, as well as require unique input modalities and interaction techniques. In attempting to resolve this issue, researchers have found that dictionary-based predictive disambiguation text entry methods are fairly efficient for text entry on devices such as mobile phones that use keypads instead of full keyboards. This type of text entry method ""guesses"" the word that a user desires by matching their sequence of keystrokes against feasible entries saved in a dictionary. However, word ambiguity, limited dictionary sizes, and large learning curves still prevent this method from being more widely adopted in many situations, and on more mobile devices. Innovative solutions to these problems, focusing on both physical keypad designs and predictive disambiguation methods, are introduced in this dissertation work. The first part of this dissertation describes a set of keypad designs which are optimized under the constraint of keeping characters in alphabetical order across keys. Designs were found that have performance close to that of unconstrained designs, while maintaining better novice usability. The second part proposes a novel predictive disambiguation method which utilizes not only word frequency information, as do most existing dictionary-based predictive disambiguation methods, but also semantic and syntactical text information to help disambiguate the user's desired words. Simulations and an empirical user study have shown improvements in text entry speed of up to 9.6% and reductions in the number of user errors of up to 21.2%. Furthermore, this dissertation presents a new error metric that is capable of revealing more information about user performance during experiments involving text entry methods. In summary, this dissertation work focused on creating and validating improved methods for text entry on mobile devices.

Document Type

Dissertation

Rights Information

Copyright 2007

Rights Holder

Jun Gong


View Full Text

Click button above to open, or right-click to save.

Share

COinS