Some relevant files:
- perceptron.py: The perceptron and kernel perceptron algorithms, written by Mathieu Blondel, October 2010
- pdc.py: Train and test deep cascades with perceptron
- compare_dumps.py: Compare best classifiers from certain dumps and choose the best one from them
- pdc_multiprocess.py: Use pdc.py with compare_dumps.py to run as multiple processes, bypassing the GIL
- split_sets.py: Split a dataset into training and testing sets
- generate_folds.py: Split a training set into three folds of training and testing sets
- crossvalidate.py: Train on the folds per a particular gamma and dataset and dump the result
- evaluate.py: Evaluate the dumps from crossvalidate.py to determine the crossvalidation results
- svm_test.py: Cross-validate for degree and C for SVM and get testing error