quantum _stock_price_simulation_mean.py

#!/usr/bin/env python
# coding: utf-8

# ## **Importing Libraries**


from tfqml import *


# ## **Choose the 8 stocks**

name_stocks = ['AAPL','ABB','ABBV','TOT','WMT','DUK','CHL','HSBC']

size = 10

# # Classical Preprocessing
for name in name_stocks:
	cp = ClassicalPreprocessing(filename=name)
	days = 30
	x_train,y_train,x_test,y_test = cp.preprocessing(days=days,flag_pca=False)


	# # Quantum Pre processing


	qp = QuantumPreprocessing(x_train,x_test)
	x_trainq,x_testq = qp.data2qubits()


	# Print a quantum circuit of our encoding


	qp.print_circuit(0)


	qubits_required = days//3
	
	list_error_train = []
	list_error_test = []
	for iteration in range(size):
		qm = QuantumModel(qubits_required)
		model_circuit, model_readout =qm.quantum_circuit([cirq.Y,cirq.X])

		# Creating our LSTM hibryd model:



		# Build the Keras model.
		model = tf.keras.Sequential([
		    # The input is the data-circuit, encoded as a tf.string
		    tf.keras.layers.Input(shape=(), dtype=tf.string),
		    # The PQC layer returns the expected value of the readout gate.
		    tfq.layers.PQC(model_circuit, model_readout),
		    tf.keras.layers.Reshape((1,10)),
		    tf.keras.layers.LSTM(units=100, return_sequences=True),
		    tf.keras.layers.Dropout(0.2),
		    tf.keras.layers.LSTM(units=100, return_sequences=True),
		    tf.keras.layers.Dropout(0.2),
		    tf.keras.layers.LSTM(units=5),
		    tf.keras.layers.Dense(units=1),
		])


		# In[8]:


		model.compile(
		    loss='mean_squared_error',
		    optimizer=tf.keras.optimizers.Adam(),   		     metrics=[tf.keras.metrics.RootMeanSquaredError()])




		qnn_history = model.fit(
		      x_trainq, y_train,
		      batch_size=32,
		      epochs=60)

		predictions_train = model.predict(x_trainq)
		list_error_train.append(mean_absolute_error(y_train, predictions_train)*100)
		
		
		predictions_test = model.predict(x_testq)
		list_error_test.append(mean_absolute_error(y_test, predictions_test)*100)
		save_data(model,x_trainq,y_train,name+"_"+str(iteration)+"_train")
		save_data(model,x_testq,y_test,name+"_"+str(iteration)+"_test")
	
	
	df = pd.DataFrame({'error_train':list_error_train, 'error_test':list_error_test })
	df.to_csv(name+'_error.csv', index=False)