nchoices=nchoices,
decay_type='threshold',
batch_train=True)
adaptive_greedy_perc = AdaptiveGreedy(deepcopy(base_algorithm),
nchoices=nchoices,
beta_prior=beta_prior,
decay_type='percentile',
decay=0.9997,
batch_train=True)
explore_first = ExploreFirst(deepcopy(base_algorithm),
nchoices=nchoices,
beta_prior=None,
explore_rounds=1500,
batch_train=True)
active_explorer = ActiveExplorer(deepcopy(base_algorithm),
nchoices=nchoices,
beta_prior=beta_prior,
batch_train=True)
adaptive_active_greedy = AdaptiveGreedy(deepcopy(base_algorithm),
nchoices=nchoices,
beta_prior=beta_prior,
active_choice='weighted',
decay_type='percentile',
decay=0.9997,
batch_train=True)
softmax_explorer = SoftmaxExplorer(deepcopy(base_algorithm),
nchoices=nchoices,
beta_prior=beta_prior,
batch_train=True)
models = [
bootstrapped_ucb, bootstrapped_ts, one_vs_rest, epsilon_greedy,
batch_train=True,
refit_buffer=50,
deep_copy_buffer=False,
random_state=3333)
adaptive_greedy_perc = AdaptiveGreedy(deepcopy(base_ols),
nchoices=nchoices,
smoothing=(1, 2),
beta_prior=None,
decay_type='percentile',
decay=0.9997,
batch_train=True,
random_state=4444)
active_explorer = ActiveExplorer(deepcopy(base_sgd),
smoothing=None,
nchoices=nchoices,
beta_prior=((3. / nchoices, 4.), 2),
batch_train=True,
refit_buffer=50,
deep_copy_buffer=False,
random_state=5555)
epsilon_greedy_nodecay = EpsilonGreedy(deepcopy(base_ols),
nchoices=nchoices,
smoothing=(1, 2),
beta_prior=None,
decay=None,
batch_train=True,
deep_copy_buffer=False,
random_state=6666)
models = [
linucb, adaptive_active_greedy, softmax_explorer, adaptive_greedy_perc,
active_explorer, epsilon_greedy_nodecay