#Find the deltas in pd (column 11) with the learners (column 4) running on the data (column 1).
# WARNING: slow- takes up to a minute. So be nice to others when running large jobs.

        printf "finding deltas (takes <60 secs) ...\n"
        nice -n 19 deltas Data=1 Rx=4 Effect=11 Fields=14 Include='logNums' ../pstats.out > $Sandbox/logNums.deltas

#Save some of  input and output of this program

        head -100 ../pstats.out            > $Safe/pstats100.deltas
        head -100 $Sandbox/logNums.deltas > $Safe/logNums100.deltas

# Generate box plots from the deltas.
# WARNING: slow- takes up to a minute. So be nice to others when running large jobs.
    
       printf "summarizing (takes <60 secs) ...\n\n"
       nice -n 19 deltaSummary $Sandbox/logNums.deltas  | malign | tee $Safe/logNums.box

# This should generate something like this:

# #rx,  mean,   sd,      n,    min,    q1, median,   q3,   max,    ------------------------0-------------------------
#   nb,  42.9, 41.6, 187132, -100.0,   0.0,   50.0, 82.4, 100.0, *************************             |         *****
#  j48,  -5.8, 43.7, 187108, -100.0, -28.5,    0.0, 11.8, 100.0, ******************        |   ***********************
# oneR, -37.1, 39.5, 187108, -100.0, -76.9,  -23.1,  0.0,   0.0,                                 ******              |

# Now apply the statistical difference tests:

      printf "\n\n"
      changes $Safe/logNums.box | malign | tee $Safe/logNums.changes

# ID, deltas(95%),  #rx,  mean,   sd,      n,    min,    q1, median,   q3,   max,    ------------------------0-------------------------
#  1,       (= 1),   nb,  42.9, 41.6, 187132, -100.0,   0.0,   50.0, 82.4, 100.0, *************************             |         *****
#  2,       (< 1),  j48,  -5.8, 43.7, 187108, -100.0, -28.5,    0.0, 11.8, 100.0, ******************        |   ***********************
#  3,       (< 2), oneR, -37.1, 39.5, 187108, -100.0, -76.9,  -23.1,  0.0,   0.0,                                 ******              |