Open davebraze opened 8 years ago
Here's a recent paper describing a hill climbing approach to fixation location adjustment. See if it's useful to us. Also see references cited there.
Palmer, C., & Sharif, B. (2016). Towards automating fixation correction for source code. Paper presented at the Proceedings of the Ninth Biennial ACM Symposium on Eye Tracking Research & Applications, Charleston, South Carolina. http://dl.acm.org/citation.cfm?id=2857544
Experiment with various classification schemes for adjusting y position (adjust_y.R), assigning particular fixations to specific lines. Possibly start with Cubist https://cran.r-project.org/web/packages/Cubist/vignettes/cubist.pdf.