softml: Semi-supervised learning

Work in progress. More details later.

How to run an experiment

cd python
python run.py

This will run the default experiment (../experiments/default.yaml).

You can use another YAML file as a parameter:

python run.py ../experiments/svn_rbf.yaml

Feel free to check out the ../experiments directory to see how to define a YAML file with experiment settings.

data: for data files used in the experiments. Some of them are available from UCL Machine Learning Repository and were not redistributed here.
experiments: files that define experiment parameters (what dataset is used, what algorithm etc.) This should help to reproduce the results for each experiment.
experiments/results: automatically generated files that contain a result of each experiment. File names are generated automatically as well: file name of the initial experiment and a time stamp.
python: my Python code to run the experiments. Under construction.

Name		Name	Last commit message	Last commit date
Latest commit History 175 Commits
data		data
experiments		experiments
generate-data		generate-data
python		python
.gitignore		.gitignore
README.md		README.md