def buildnetwork(self):
model = Sequential()
model.add(lstm(20, dropout=0.2,input_shape = (self.seq_len, self.n_feature)))
model.add(Dense(1, activation=None))
model.compile(loss='mean_squared_error', optimizer=Adagrad(lr=0.002,clipvalue=10), metrics=['mean_squared_error'])
return model
def buildModel(shape, lstm_layers, lstm_units, dense_layers, dense_units):
print(shape)
model = Sequential()
# LSTM layers
for l in range(lstm_layers):
if l == 0:
this_shape = (shape[1], shape[2])
else:
this_shape = (lstm_units, 1)
if l == lstm_layers - 1:
this_return = False
else:
this_return = True
model.add(lstm(lstm_units, input_shape=this_shape, return_sequences=this_return))
model.add(Activation('relu'))
# Dense layers
for l in range(dense_layers):
model.add(Dense(dense_units))
model.add(Activation('relu'))
model.add(Dropout(0.1))
model.add(Dense(1))
model.compile(loss="mse", optimizer="adam", metrics=['mae'])
model.summary()
return model
def buildnetwork(self):
model = Sequential()
model.add(
lstm(20, dropout=0.2, input_shape=(self.seq_len, self.n_feature)))
model.add(Dense(1, activation=None))
model.compile(loss='mean_squared_error',
optimizer=Adagrad(lr=0.002, clipvalue=10),
metrics=['mean_squared_error'])
return model
def make_default_model(self):
self.model.add(
lstm(128, input_shape=(self.input_shape[0], self.input_shape[1])))
self.model.add(Dropout(0.5))
self.model.add(Dense(32, activation='relu'))
self.model.add(Dense(16, activation='tanh'))