def __init__(self,
env,
optim=Adam,
policy_lr=0.001,
value_lr=0.001,
policy_hidden_size=[32],
value_hidden_size=[32],
gamma=0.9,
batch_size=5000,
epochs=50,
update_every=50,
render=False):
self.env = env
self.batch_size = batch_size
self.render = render
self.epochs = epochs
self.gamma = gamma
self.update_every = update_every
obs_size = np.prod(env.observation_space.shape)
action_size = env.action_space.n
self.policy_mlp = MLP([obs_size] + policy_hidden_size + [action_size])
self.policy_optim = optim(self.policy_mlp.parameters(), lr=policy_lr)
self.value_mlp = MLP([obs_size] + value_hidden_size + [1])
self.value_optim = optim(self.value_mlp.parameters(), lr=value_lr)
def __init__(self,
env,
optim=Adam,
policy_lr=0.01,
value_lr=0.1,
policy_hidden_size=[32],
value_hidden_size=[32],
batch_size=5000,
render=False):
self.env = env
self.batch_size = batch_size
self.render = render
obs_size = np.prod(env.observation_space.shape)
action_size = env.action_space.n
self.policy_mlp = MLP([obs_size] + policy_hidden_size + [action_size])
self.policy_optim = optim(self.policy_mlp.parameters(), lr=policy_lr)
self.value_mlp = MLP([obs_size] + value_hidden_size + [1])
self.value_optim = optim(self.value_mlp.parameters(), lr=value_lr)
def __init__(self,
env,
optim=Adam,
lr=0.01,
hidden_size=[64],
batch_size=5000,
n_episodes=2000,
render=False):
self.env = env
self.batch_size = batch_size
self.n_episodes = n_episodes
self.lr = lr
self.render = render
obs_size = np.prod(env.observation_space.shape)
action_size = env.action_space.n
self.mlp = MLP([obs_size] + hidden_size + [action_size])
self.optim = optim(self.mlp.parameters(), lr=lr)
if "NRF" in list(data):
data.pop("NRF")
if "POSTCR" in list(data):
data.pop("POSTCR")
if "OpID" in list(data):
data.pop("OpID")
if "PatID" in list(data):
data.pop("PatID")
if "DOA" in list(data):
data.pop("DOA")
y = data.pop("HAEMOFIL")
ros = RandomOverSampler(random_state=1)
scaler = scale()
scaler.fit(data)
print("full")
param = {'layers':[2,5], 'nodes':[5,10], 'dropout':[0.4,0.8], 'epochs':[50]}
gsearch = GridSearchCV(estimator = MLP(),
param_grid = param, scoring='roc_auc', iid=False, cv=rkf_search, verbose=2)
gsearch.fit(scaler.transform(data.values), y.values)
clf = gsearch.best_estimator_
pd.DataFrame(gsearch.cv_results_).to_csv("output/HF/MLPfull.csv")
output = cross_validate(clf, scaler.transform(data.values), y.values, scoring=metrics,cv=rkf, verbose=2,return_train_score=True)
pd.DataFrame(output).to_csv('output/HF/performanceMLPfull.csv')