def metabolite_combined_dataset():
"""
Runs the combined dataset of OAT1-OAT3-OATP.
The GA will receive all the features together.
Then, the random forest on top will use those features and get the
accuracy on the dataset using 10-fold-cross validation.
:return: None.
"""
num_epochs = prompt_num_epochs()
dl = DataLoaderMetabolite()
train_data, train_labels, header = dl.load_oat1_3_p_combined()
FeatureEngineering.acc_function = feature_eng_err_metab_comb
algo = BaseGA('./configs/metabolite_FE.config',
checkpoint_prefix='FE_metab_cb_')
FeatureEngineering.run_session(algo, header, num_epochs)
def metabolite_combined_dataset():
"""
Runs the small dataset of OAT1-OAT3.
The GA will receive all the features together and then use leave one
out to find its accuracy over given number of epochs.
:return: None.
"""
num_epochs = prompt_num_epochs()
dl = DataLoaderMetabolite()
train_data, train_labels, header = dl.load_oat1_3_p_combined()
algo = BaseGA('./configs/metabolite_MULTI.config',
checkpoint_prefix='GA_metab_cb_')
FeatureSelectionGA.acc_function = feature_sel_err_metab_comb
FeatureSelectionGA.run_session(algo, header, num_epochs)
def metabolite_small_dataset():
"""
Runs the small dataset of OAT1-OAT3.
The GA will receive all the features together.
Then, the random forest on top will use those features and get the
accuracy on the dataset using leave one out.
:return: None.
"""
num_epochs = prompt_num_epochs()
dl = DataLoaderMetabolite()
train_data, train_labels, header = dl.load_oat1_3_small()
FeatureEngineering.acc_function = feature_eng_err_metab_small
algo = BaseGA('./configs/metabolite_FE.config',
checkpoint_prefix='FE_metab_sm_')
FeatureEngineering.run_session(algo, header, num_epochs)