Question

3) Create a user defined function for the given Python program. Importing data into Python from filename.csv • Use the same parameters and variables as defined in the following Python programs. • Run your programs 3- Gradient descent method error_sq_list = [] # Loss mean square error list bo, b1 = 0,0 # Initialization of the parameters alpha = 10e-1 # Learning rate num_iterations = 100 # Number of iterations/replications len_samples = len(sales) # Number of samples = m samples = range(len_samples) # List of samples index count = 0 while count < num_iterations: dbo, db1 = 0,0 # The derivative of the parameters with respect to the loss function L = dL/db error_sq = 0 # error_sq = (predicted sales - actual sales)^2 for i in samples: predict = b0 + b1 * tv_ad[i] # prediction => prediction of sales = b0hat + b1hat * tv_add predict_minus_actual = predict - sales[i] db0 += predict_minus_actual db1 += predict_minus_actual * tv_ad[i] error_sq += predict_minus_actual**2 # Compute square error and accumulate bo = bo - alpha * (dbo / len_samples) b1 = b1 - alpha * (db1 / len_samples) error_sq_list.append((error_sq / len_samples)) count += 1 print ("\n Gradient Descent Method: sales = ", round(b0,3), " + ", round(b1,3), " * tv_ad")

          3) Create a user defined function for the given Python program. Importing data
into Python from filename.csv
• Use the same parameters and variables as defined in the following Python
programs. • Run your programs
3- Gradient descent method
error_sq_list = [] # Loss mean square error list
bo, b1 = 0,0 # Initialization of the parameters
alpha = 10e-1 # Learning rate
num_iterations = 100 # Number of iterations/replications
len_samples = len(sales) # Number of samples = m
samples = range(len_samples) # List of samples index
count = 0
while count < num_iterations:
dbo, db1 = 0,0 # The derivative of the parameters with respect to the loss function L = dL/db
error_sq = 0 # error_sq = (predicted sales - actual sales)^2
for i in samples:
predict = b0 + b1 * tv_ad[i] # prediction => prediction of sales = b0hat + b1hat * tv_add
predict_minus_actual = predict - sales[i]
db0 += predict_minus_actual
db1 += predict_minus_actual * tv_ad[i]
error_sq += predict_minus_actual**2 # Compute square error and accumulate
bo = bo - alpha * (dbo / len_samples)
b1 = b1 - alpha * (db1 / len_samples)
error_sq_list.append((error_sq / len_samples))
count += 1
print ("\n Gradient Descent Method: sales = ", round(b0,3), " + ", round(b1,3), " * tv_ad")

3) Create a user defined function for the given Python program. Importing data
into Python from filename.csv
• Use the same parameters and variables as defined in the following Python
programs. • Run your programs
3- Gradient descent method
errorsqlist = [] # Loss mean square error list
bo, b1 = 0,0 # Initialization of the parameters
alpha = 10e-1 # Learning rate
numiterations = 100 # Number of iterations/replications
lensamples = len(sales) # Number of samples = m
samples = range(lensamples) # List of samples index
count = 0
while count < numiterations:
dbo, db1 = 0,0 # The derivative of the parameters with respect to the loss function L = dL/db
errorsq = 0 # errorsq = (predicted sales - actual sales)^2
for i in samples:
predict = b0 + b1 * tvad[i] # prediction => prediction of sales = b0hat + b1hat * tvadd
predictminusactual = predict - sales[i]
db0 += predictminusactual
db1 += predictminusactual * tvad[i]
errorsq += predictminusactual**2 # Compute square error and accumulate
bo = bo - alpha * (dbo / lensamples)
b1 = b1 - alpha * (db1 / lensamples)
errorsqlist.append((errorsq / lensamples))
count += 1
print ("Gradient Descent Method: sales = ", round(b0,3), " + ", round(b1,3), " * tvad")

Added by Thomas M.

Question

Please give Ace some feedback