search

XueqiYang / static_analysis

0 stars 0 forks source link

from rahul. RF+CFS, GHOST #2

Open timm opened 2 years ago

timm commented 2 years ago

Preliminary Results

RF + CFS

                 acc      prec       rec        f1        pf       auc
ant          0.600000  0.600000  1.000000  0.750000  1.000000  0.500000
cassandra    0.857143  0.800000  1.000000  0.888889  0.333333  0.833333
commons      0.200000  0.153846  0.285714  0.200000  0.846154  0.219780
derby        0.733333  0.733333  1.000000  0.846154  1.000000  0.500000
jmeter       0.538462  0.541958  1.000000  0.700000  0.833333  0.505952
lucene-solr  0.888889  0.909091  0.909091  0.909091  0.142857  0.883117
tomcat       0.590643  0.655844  0.669903  0.663462  0.529412  0.567961

GHOST

                accuracy         f1        pd            pf      prec       auc
               ----------  --------  --------      --------  --------  -------- 
ant          0.6          0.75         1             1       0.6       0.5
cassandra   0.857143      0.888889         1      0.333333       0.8  0.833333
derby       0.688889           0.8  0.848485          0.75  0.756757  0.549242
jmeter      0.611111      0.758621         1             1  0.611111       0.5
tomcat      0.619883      0.679803   0.669903    0.455882       0.69   0.60701
timm commented 2 years ago

@rahul: please report things in plain text, justifed columns, 3 sig figs at most. please

  1. is this from sherry' 84 data sets? you need seperate results for each of the 4 groups
  2. where is commentary on the above. e.g i would say the PFs are awful. what can you do about this?
yrahul3910 commented 2 years ago

@timm no, these are from the original reimplemented_2016_manual files. will redo with the 84 datasets.

timm commented 2 years ago

@yrahul3910 : k