Skip to content

Latest commit

 

History

History
93 lines (66 loc) · 2.6 KB

README.md

File metadata and controls

93 lines (66 loc) · 2.6 KB

eyeFixationData

Introduction:

The following is an overview of the R code associated with the article "Markov models for ocular fixation locations in the presence and absence of colour" by Adam B Kashlak, Eoin Devane, Helge Dietert, and Henry Jackson from the University of Cambridge. Any questions or comments can be directed at Adam B Kashlak ([email protected]).

Note that most all of the functions in the R code begin with the prefix 'rss'. For the curious, the reason is because most of this code was written in 2015 RSS Statistical Analytics Challenge. (see https://rsschallenge.wordpress.com/)

Files Included:

The R code: eyeMarkovCode4JRSSC.r The data table: data.csv The directory of images: images/

Data:

To access the data from the R code, the user must set the variable 'dataDir' at the top of the .r file. This currently defaults to the current working directory "./". The data table is contained in ./data.csv. The image files can be found in the subdirectories of ./images/. For the sake of statistics, the image files are not necessary. However, if they are absent, then many of the display functions in the R code will fail to run correctly.

Data is available on GitHub at https://github.com/cachelack/eyeFixationData

Code Examples:

The R code is a bit cumbersome as it has not been cleaned into a nice library. Here are some examples of functions to run to reproduce results from the article.

  • To display photos with plotted fixations with 'photo_number' in {1,2,...,60} and 'type' in {'normal','abnormal','grayscale'}

rss_plotImageData2( photo_number,type )

  • To extract some subset of the data.csv table with image imag in {1,...,60} condition type in {"normal","grayscale","abnormal"} subject subj in {1,...,10} fixation fixt in {1,...,26}

rss_getData <- function( imag=F, type=F, subj=F, fixt=F )

  • Similarly, to extract data about the transitions, which is the Euclidean distance between successive fixations.

rss_getSacData <- function( imag=F, type=F, subj=F, fixt=F )

  • To recreate Figure 6,

rss_plotDurations()

  • From Section 2 of the article: To compute Bayes factors for model selection and choose an optimal clustering, use rss_computeAllBayes( ). For example, so analyze the first 2 photos over all three colour schemes, run

rss_computeAllBayes( images=1:2 )

  • This function basically calls rss_mmBayesFactor( ) on all of the specified images. This can be called individually as follows

out = rss_mmBayesFactor( 1, 'normal' )

It produces all of the necessary output for further analysis as in the ROC curve from Figure 3