algorithm/Python
1
0
Fork 0
mirror of https://github.com/TheAlgorithms/Python.git synced 2026-03-13 09:50:19 +08:00
Code Issues Packages Projects Releases Wiki Activity

psf/black code formatting (#1277)

Browse Source
This commit is contained in:
William Zhang
2019-10-05 01:14:13 -04:00
committed by Christian Clauss
parent 07f04a2e55
commit 9eac17a408
291 changed files with 6014 additions and 4571 deletions
Download Patch File Download Diff File Expand all files Collapse all files

46
machine_learning/gradient_descent.py
View File

@@ -4,21 +4,28 @@ Implementation of gradient descent algorithm for minimizing cost of a linear hyp
import numpy
# List of input, output pairs
train_data = (((5, 2, 3), 15), ((6, 5, 9), 25),
((11, 12, 13), 41), ((1, 1, 1), 8), ((11, 12, 13), 41))
train_data = (
((5, 2, 3), 15),
((6, 5, 9), 25),
((11, 12, 13), 41),
((1, 1, 1), 8),
((11, 12, 13), 41),
)
test_data = (((515, 22, 13), 555), ((61, 35, 49), 150))
parameter_vector = [2, 4, 1, 5]
m = len(train_data)
LEARNING_RATE = 0.009
def _error(example_no, data_set='train'):
def _error(example_no, data_set="train"):
"""
:param data_set: train data or test data
:param example_no: example number whose error has to be checked
:return: error in example pointed by example number.
"""
return calculate_hypothesis_value(example_no, data_set) - output(example_no, data_set)
return calculate_hypothesis_value(example_no, data_set) - output(
example_no, data_set
)
def _hypothesis_value(data_input_tuple):
@@ -32,7 +39,7 @@ def _hypothesis_value(data_input_tuple):
"""
hyp_val = 0
for i in range(len(parameter_vector) - 1):
hyp_val += data_input_tuple[i]*parameter_vector[i+1]
hyp_val += data_input_tuple[i] * parameter_vector[i + 1]
hyp_val += parameter_vector[0]
return hyp_val
@@ -43,9 +50,9 @@ def output(example_no, data_set):
:param example_no: example whose output is to be fetched
:return: output for that example
"""
if data_set == 'train':
if data_set == "train":
return train_data[example_no][1]
elif data_set == 'test':
elif data_set == "test":
return test_data[example_no][1]
@@ -75,7 +82,7 @@ def summation_of_cost_derivative(index, end=m):
if index == -1:
summation_value += _error(i)
else:
summation_value += _error(i)*train_data[i][0][index]
summation_value += _error(i) * train_data[i][0][index]
return summation_value
@@ -85,7 +92,7 @@ def get_cost_derivative(index):
:return: derivative wrt to that index
Note: If index is -1, this means we are calculating summation wrt to biased parameter.
"""
cost_derivative_value = summation_of_cost_derivative(index, m)/m
cost_derivative_value = summation_of_cost_derivative(index, m) / m
return cost_derivative_value
@@ -99,11 +106,16 @@ def run_gradient_descent():
j += 1
temp_parameter_vector = [0, 0, 0, 0]
for i in range(0, len(parameter_vector)):
cost_derivative = get_cost_derivative(i-1)
temp_parameter_vector[i] = parameter_vector[i] - \
LEARNING_RATE*cost_derivative
if numpy.allclose(parameter_vector, temp_parameter_vector,
atol=absolute_error_limit, rtol=relative_error_limit):
cost_derivative = get_cost_derivative(i - 1)
temp_parameter_vector[i] = (
parameter_vector[i] - LEARNING_RATE * cost_derivative
)
if numpy.allclose(
parameter_vector,
temp_parameter_vector,
atol=absolute_error_limit,
rtol=relative_error_limit,
):
break
parameter_vector = temp_parameter_vector
print(("Number of iterations:", j))
@@ -111,11 +123,11 @@ def run_gradient_descent():
def test_gradient_descent():
for i in range(len(test_data)):
print(("Actual output value:", output(i, 'test')))
print(("Hypothesis output:", calculate_hypothesis_value(i, 'test')))
print(("Actual output value:", output(i, "test")))
print(("Hypothesis output:", calculate_hypothesis_value(i, "test")))
if __name__ == '__main__':
if __name__ == "__main__":
run_gradient_descent()
print("\nTesting gradient descent for a linear hypothesis function.\n")
test_gradient_descent()