Q. How do I run the script?
A. Execute com.ase.extn.cart.cart.py
Q. What parameters do I need to set in the script?
A. Edit com.ase.extn.cart.cart.py to modify following variables -
- strategy = 'projective'|'progressive'
- system = 'all'|'apache'|'bc'|'bj'|'llvm'|'sqlite'|'x264'
Q. How can I interpret the results?
A. For progressive sampling, individual output files are created under data/ouput for each system. These files contain a mapping between sample size and fault rate.
For projective sampling, results are displayed in the console in the following format:
System-id : <system under execution>
Size of lambda set: <size of the data points used for the projected learning curve>
------------------------------------------------------------------
{<projected curve> : [<correlation value>, <optimal sample size>, <accuracy of prediction with optimal sample size>, <standard deviation>]}
Q. Where do I feed in the input files?
A. Input files are read from data/input
Q. Where can I get further details?
A. http://gsd.uwaterloo.ca/sites/default/files/PID3840471.pdf