def wisconsin_breast_experiment():
ExperimentSetup.wisconsin_breast_cancer_data({
'method': 'full',
'sparse_factor': 1.0,
'run_id': 1,
'log_level': logging.DEBUG
})
def abalone_experiment():
ExperimentSetup.abalone_data({
'method': 'full',
'sparse_factor': 1.0,
'run_id': 1,
'log_level': logging.DEBUG
})
def mining_experiment():
ExperimentSetup.mining_data({
'method': 'mix1',
'sparse_factor': 1.0,
'run_id': 1,
'log_level': logging.DEBUG
})
def boston_experiment():
ExperimentSetup.boston_data({
'method': 'mix2',
'sparse_factor': 0.8,
'run_id': 3,
'log_level': logging.DEBUG
})
def sarcos_experiment():
ExperimentSetup.sarcos_data({'method': 'full',
'sparse_factor': 0.04,
'run_id': 0,
'log_level': logging.DEBUG,
'n_thread': 15,
'partition_size': 2000,
# 'image': '../results/all/'
})
def main(args):
algorithm = args['algo']
with tf.Session() as sess:
print(args['env'])
exp_setup = ExperimentSetup(algorithm, args['env'], sess, args['random_seed'])
exp_setup.setup_experiment(args)
train_experiment(algorithm, exp_setup)
def sarcos_experiment():
ExperimentSetup.sarcos_data({'method': 'full',
'sparse_factor': 0.04,
'run_id': 0,
'log_level': logging.DEBUG,
'n_thread': 15,
'partition_size': 2000,
# 'image': '../results/all/'
})
def USPS_experiment():
ExperimentSetup.USPS_data({
'method': 'full',
'sparse_factor': 0.1,
'run_id': 1,
'log_level': logging.DEBUG
})
def mnist_binary_inducing_experiment():
ExperimentSetup.MNIST_binary_inducing_data({'method': 'full',
'sparse_factor': 200. / 60000,
'run_id': 1,
'log_level': logging.DEBUG,
'n_thread': 8,
'partition_size': 1000,
# 'image': '../results/mnist_1/'
})
def creep_experiment():
ExperimentSetup.creep_data({'method': 'full', 'sparse_factor': 1.0, 'run_id': 1, 'log_level': logging.DEBUG})
def wisconsin_breast_experiment():
ExperimentSetup.wisconsin_breast_cancer_data(
{'method': 'full', 'sparse_factor': 1.0, 'run_id': 1, 'log_level': logging.DEBUG})
def boston_experiment():
ExperimentSetup.boston_data({'method': 'mix2', 'sparse_factor': 0.8, 'run_id': 3, 'log_level': logging.DEBUG})
", $\Theta = \{\stackrel{6.5}{0}\}$")[0], )
tar_labels.add(target_to_cls_dict[i])
ax.legend(handles=patches, fontsize=18, loc='best')
if show:
plt.show()
return fig
if __name__ == "__main__":
env = 'phantom2'
key = list(pca_sample_numb_change.keys())[0]
exp_setup = ExperimentSetup(data_path[env],
baseline="./../data/baseline.mat")
exp_setup.set_class_names(properties_exp_state[env])
exp_metric_data, exp_fig_data = exp_setup.run_experiments(
which_type='properties',
which_clustering='k_means',
show=SHOW,
save_local=SAVE_LOCAL,
save_global=SAVE_GLOBAL,
where=env,
resolution=RESOLUTION,
format=SAVE_FORMAT)
# -------------------------------------------------------------------------
# --------------------------- FIGURES -----------------------------------
# -------------------------------------------------------------------------
}
if __name__ == "__main__":
cluster_numbers = list(class_types.keys())
exp_metric_data = dict()
exp_fig_data = dict()
exp_setup = dict()
env = list(properties_exp_state.keys())[0]
data_path = data_path[list(data_path.keys())[0]]
for clst_no in cluster_numbers:
exp_setup[clst_no] = ExperimentSetup(data_path,
baseline="./../data/baseline.mat")
exp_setup[clst_no].set_class_names(properties_exp_state[env])
exp_setup[clst_no].set_class_type(class_types[clst_no])
exp_metric_data[clst_no], _ = exp_setup[clst_no].run_experiments(
which_type='properties',
which_clustering='k_means',
show=SHOW,
save_local=SAVE_LOCAL,
save_global=SAVE_GLOBAL,
where=env,
resolution=RESOLUTION,
format=SAVE_FORMAT)
############# BAR PLOT FIGURE SILHOUETTE ##################
clst_no_fig = plt.figure(figsize=(18, 10))
from experiment_setup import ExperimentSetup
task2 = frozenset(
{frozenset({'Sphere', 'HSphere', 'HCube'}),
frozenset({'Cube'})})
task4 = frozenset(
{frozenset({'Sphere', 'HSphere'}),
frozenset({'HCube', 'Cube'})})
task5 = frozenset(
{frozenset({'Sphere', 'Cube'}),
frozenset({'HCube', 'HSphere'})})
exp_setup = ExperimentSetup("./../data/skin_experiments.mat",
best_pressure_time=140)
# exp_setup.run_experiment(experiment_name='3mil', which_clustering='k_means', task=task4, show=True, save=True)
exp_setup.run_experiment(experiment_name='6mil',
which_clustering='k_means',
task=task2,
show=True,
save=True)
# exp_setup.run_experiment(experiment_name='10mil', which_clustering='k_means', task=task5, show=True, save=True)
'no SS': ['1-1', '2-1', '3-1', '4-1', '2-2', '3-2', '4-2', '1-3', '3-3', '4-3', '2-4', '3-4', '4-4'],
'no SD': ['1-1', '2-1', '3-1', '4-1', '1-2', '2-2', '3-2', '4-2', '1-3', '2-3', '3-3', '4-3', '1-4'],
}
data_path = {
'phantom1': "./../data/presence_exp.mat",
'phantom2': "./../data/properties_exp.mat"
}
exp_metric_data = dict()
exp_fig_data = dict()
exp_setup = dict()
env = 'phantom2'
for key in pca_sample_numb_change.keys():
exp_setup[key] = ExperimentSetup(data_path[env], baseline="./../data/baseline.mat")
exp_setup[key].set_class_names(properties_exp_state[env])
exp_metric_data[key], _ = exp_setup[key].run_experiments(
which_type='presence',
which_clustering='k_means',
show=SHOW,
save_local=SAVE_LOCAL,
save_global=SAVE_GLOBAL,
where=env,
resolution=RESOLUTION,
format=SAVE_FORMAT,
# bad extra parameters
pca_sample_numb_change=pca_sample_numb_change[key]
)
# -------------------------------------------------------------------------
def mining_experiment():
ExperimentSetup.mining_data({'method': 'mix1', 'sparse_factor': 1.0, 'run_id': 1, 'log_level': logging.DEBUG})
