Final Project CSC 591

How to run

• Install python 3.9 or newer
• Install requirements: pip install -r requirements.txt
• Change directory to src using cd src
• Run python main.py to generate table for the auto2.csv file
• To run all datasets in the data folder at once, please do sh generate_output.sh
• Run python main.py --help to view possible configuration values

project: multi-goal semi-supervised algorithms

USAGE: python3 main.py [OPTIONS] [-g ACTIONS]
  
OPTIONS:
  -b  --bins    initial number of bins       = 16
  -c  --cliffs  cliff's delta threshold      = .147
  -d  --d       different is over sd*d       = .35
  -F  --Far     distance to distant          = .95
  -g  --go      start-up action              = nothing
  -f  --file    file to generate table of    = etc/data/new_SSM_for_FebStudy.csv
  -h  --help    show help                    = false
  -H  --Halves  search space for clustering  = 315
  -m  --min     size of smallest cluster     = .5
  -M  --Max     numbers                      = 512
  -p  --p       dist coefficient             = 2
  -r  --rest    how many of rest to sample   = 50
  -x  --bootstrap   number of samples to bootstrap   = 512
  -R  --Reuse   child splits reuse a parent pole = false
  -n  --nTimes   number of iterations to run  = 20
  -m  --mSeed    true means seed isn't randomized  = false
  -o  --conf        confidence interval      = 0.05
  -s  --seed    random number seed           = 937162211
  -h  --help    show help                    = false
  -w  --wColor      output with color        = true
  -w1 --width    the width for the tile      = 40
  -f1 --Fmt     the format of the tile       = {:.2f}
  -c1 --cohen   effect-size mthod            = .35

For Repgrids report, please see docs/repertory_grids.md.

Project Structure

• Data files are located at data
• Output files are located at out
• Python files are located at src
• A script is used to generate all output and stored them to etc/out/data_file.out

Authors

• Leo Hsiang
• Ashrita Ramaswamy
• Sanjana Cheerla
• Natalie Kraft

Support

Contributions, issues, and feature requests are welcome!

Help

Email any queries to the contributors