@return: The score. We are trying to minimize this score.
"""
global input_data
global output_data
# Calculate the actual output of the polynomial with the specified coefficients.
actual_output = []
for input_data in training_input:
x = input_data[0]
output_data = poly(coeff, x)
actual_output.append(output_data)
# Build the training set.
training_input, training_ideal = build_training_set()
# Extract the input and ideal training.
training_input = np.array(training_input)
training_ideal = np.array(training_ideal)
# Starting point for coefficients.
x0 = [0, 0, 0]
# Perform the train.
train = TrainGreedRandom(-10, 10)
train.stop_score = 100
train.display_iteration = True
result = train.train(x0, score_funct)
# Evaluate the polynomial.
print("Final polynomial")
print_poly(result)
@return: The MSE error.
"""
# Setup the long-term memory that we would like to test.
network.copy_memory(x)
# Present all inputs to the network and accumulate the output for each.
actual_output = []
for input_data in training_input:
output_data = network.compute_regression(input_data)
actual_output.append(output_data)
# Compare the actual output with the ideal expected output and calculate the MSE error.
return ErrorCalculation.mse(np.array(actual_output), training_ideal)
# Use the initial long term memory of the network as the starting state.
x0 = list(network.long_term_memory)
# Train with greedy random.
train = TrainGreedRandom(-1, 1)
train.display_iteration = True
train.max_iterations = 100000
train.stop_score = 0.05
result = train.train(x0, score_funct)
# Copy the final trained long-term memory to the network so we can use it for evaluation.
network.copy_memory(result)
# Display the output for the XOR. XOR will not be trained perfectly. You should see that the (0,1) and (1,0) inputs
# are both close to 1.0, whereas the (1,1) and (0,0) are close to 0.0.
for input_data in training_input:
output_data = network.compute_regression(input_data)
print(str(input_data) + " -> " + str(output_data))
@param x: The long term memory that we are to score.
@return: The MSE error.
"""
# Setup the long-term memory that we would like to test.
network.copy_memory(x)
# Present all inputs to the network and accumulate the output for each.
actual_output = []
for input_data in training_input:
output_data = network.compute_regression(input_data)
actual_output.append(output_data)
# Compare the actual output with the ideal expected output and calculate the MSE error.
return ErrorCalculation.mse(np.array(actual_output), training_ideal)
# Use the initial long term memory of the network as the starting state.
x0 = list(network.long_term_memory)
# Train with greedy random.
train = TrainGreedRandom(-1, 1)
train.display_iteration = True
train.max_iterations = 100000
train.stop_score = 0.05
result = train.train(x0, score_funct)
# Copy the final trained long-term memory to the network so we can use it for evaluation.
network.copy_memory(result)
# Display the output for the XOR. XOR will not be trained perfectly. You should see that the (0,1) and (1,0) inputs
# are both close to 1.0, whereas the (1,1) and (0,0) are close to 0.0.
for input_data in training_input:
output_data = network.compute_regression(input_data)
print(str(input_data) + " -> " + str(output_data))
global input_data
global output_data
# Calculate the actual output of the polynomial with the specified coefficients.
actual_output = []
for input_data in training_input:
x = input_data[0]
output_data = poly(coeff, x)
actual_output.append(output_data)
return ErrorCalculation.sse(np.array(actual_output), training_ideal)
# Build the training set.
training_input, training_ideal = build_training_set()
# Extract the input and ideal training.
training_input = np.array(training_input)
training_ideal = np.array(training_ideal)
# Starting point for coefficients.
x0 = [0, 0, 0]
# Perform the train.
train = TrainGreedRandom(-10, 10)
train.stop_score = 100
train.display_iteration = True
result = train.train(x0, score_funct)
# Evaluate the polynomial.
print("Final polynomial")
print_poly(result)