Console Wfccm

Command line aspect of Wfccm. Limited capability extended.

Todo

• Add support for Shuo's preprocessed data - log transformation not applicable
• Add support for DataGrouping.MultiLoad?
  • Simplify the DataGrouping file needed
  • Must include the InformationSet
• Compile and test using Mono on GNULinux

Use

• wfccm2 -train <-generate> -d dataSet <-l logdataset> -g groupingfile -f filterfile <-r runname>
  • <scores.txt>, means.txt, train_output.txt
• wfccm2 -test -d dataSet -f filterfile -s scorefile -m meansfile <-g grouping> <-r runname> *test_output.txt

Command Line options

• -train : Indicates that a training type analysis in going to be done
• -test : Indicates that a testing type analysis in going to be done
• <-generate> : Indicates that the program needs to generate scores
• -d dataSet : path to the data set
• <-l logdataset> : Path to the log dataset
• -g groupingfile : path to the grouping file
• -f filterfile : path to the filter criteria file
• -s scorefile : path to the score file
• -m meansfile : path to the distance output means.
• <-r runname> : optional run name that will be added to the output names.
• scores.txt : score output file.
• means.txt : Means output file.
• train_output.txt : Training output
• test_output.txt : Testing output

Extended Functionality needed

1. DataGrouping.Write
2. DataGrouping.Read
3. CriteriaSet.Write
4. CriteriaSet.Read

Defaults

• Scores to be generated:
  • Sam (sam, asam)
  • Wga (wga)
  • Info (info)
  • TTest (tvalue, prob_t)
  • Fisher (fvalue, prob_f)
  • Wilcox (cvalue, prob_c)
• Function: (1 - info)*(numPass / 6)*(asam_std + tvalue_std + fvalue_std + cvalue_std)
• Sign: tvalue will set fvalue, cvalue and asam