def svm_annealing(args):
'''
Function starts simulated annealing to find out ideal parameters for SVM
classifier with given data.
'''
t = SVMTest()
t.run_annealing(n_fold_cv=args.n_fold_cv, kernel=args.kernel)
def _thread_svm(db_file, count, max_token_size, n_fold_cv, kernel):
from src.svm.svm_test import SVMTest
t = SVMTest()
# load data
t.regenerate_data(dbfile=db_file, count=count,
max_token_size=max_token_size)
# run simulated annealing
state, energy = t.run_annealing(n_fold_cv=n_fold_cv, kernel=kernel)
# run test with optimal parameters
result = t.run(c=state[1], param=state[0], n_fold_cv=n_fold_cv,
kernel=kernel)
# return results
return {'type':'svm', 'result': result, 'state':state, 'emergy':energy}