|
Replicating Reichle & Laurent (2006)
In Reichle & Laurent (2006), we instantiated some very simple assumptions about saccadic eye movements in order to observe the kinds of eye movement behavior that emerge when an agent is given the task to read as efficiently as possible. Requirements
When you unzip RL2006.zip, you will find the following directory structures: Sim1/ (no saccadic error) 01 02 ... 20 Sim2/ (saccadic error) 01 02 ... 20 "); ?> Each directory contains a reading.prop file which specifies settings for a single-sentence simulation. To run a simulation, place RLFramework.jar on your CLASSPATH, change into one of the subdirectories, and execute the simulator. For example: $ export CLASSPATH=~/Desktop/RLFramework.jar $ cd Sim1/01 $ java org.eyelanguage.rl.ReadingMain >& log.txt "); ?> If you get an OutOfMemoryError, increase the memory that Java is permitted to use: $ java -Xmx700m org.eyelanguage.rl.ReadingMain >& log.txt "); ?> The output (which is stored to log.txt) will contain the sequences of actions selected by each agent. For example, "ReqSacc3" means that the agent requested a saccade of 3 character spaces at that point in time. In order to convert the output into a format suitable for analysis in another program such as Excel, you may use: $ java org.eyelangauge.rl.analysis.LogAnalyzer log.txt > excelready.txt "); ?> |