Regression to a linear lower bound with outliers: An exponentially modified Gaussian noise model
Published in EURASIP EUSIPCO, 2019
Recommended citation: Gori, Julien, and Olivier Rioul. "Regression to a linear lower bound with outliers: An exponentially modified Gaussian noise model." 2019 27th European Signal Processing Conference (EUSIPCO). IEEE, 2019. https://hal.archives-ouvertes.fr/hal-02191051/document
In our 2018 TOCHI and 2017 Interact paper we had proposed to fit Fitts’ law on the best performing points of a scatter plot in the (MT,ID) plane. However, there is no existing method for this type of fit. In this work, we explain a method that precisely achieves this type of fitting i.e. esimating a linear lower bound of a scatter plot, when some points may cheat below that lower bound.