logs directory contains measurement data and learned performance-influence models
for each subject system:

     <subject_system_name>_sosym.a - SPLConqueror automation files.
     These are used to learn performance-influence models using SPLConqueror.

     Each directory of subject system contains:
     	  all_measurements.xml - performance measurements in an SPLConqueror format
	  feature_model.xml - variability model in an SPLConqueror format
	  feature_model.guidsl - variability model in a GUIDSL format 
	  feature_model.png - a feature diagram representation of the variability model
	  sosym_stdout - log file with learned performance-influence models

notebooks/results_analysis.ipynb is a jupyter notebook in Python 3 that analyses the data from logs directory and generates plots.

noise directory contains scripts for generating plots describing influence of measurement errors on the performance-influence model accuracy (see supplementary website http://fosd.de/tradoffs/). The corresponding measurements data is in logs/sanity_check (structured the same way as logs directory).