Unlimiting the Dual Gaussian Distribution Model to Predict Touch Accuracy in On-screen-start Pointing Tasks

Anonymous

Keywords: Dual Gaussian distribution model, touchscreens, finger input, pointing, graphical user interfaces

Abstract:

The dual Gaussian distribution hypothesis has been utilized to predict the success rate of target acquisition in finger touching.
Bi and Zhai limited the applicability of their success-rate prediction model to off-screen-start pointing.
However, we found that their doing so was theoretically over-limiting and their prediction model could also be used to on-screen-start pointing operations.
We discuss the reasons why and empirically validate our hypothesis in a series of four experiments with various target sizes and distances.
Bi and Zhai's model showed high prediction accuracy in all the experiments, with 10% prediction error at worst.
Our theoretical and empirical justifications will enable designers and researchers to use a single model to predict success rates regardless of whether users mainly perform on- or off-screen-start pointing and automatically generate and optimize UI items on apps and keyboards.