CODEWITHSUNDEEP
pythonkerasdeep-learningstandard-deviation
sharee
sharee

Reputation: 47

Summation of two normal distribution in Keras

I have a question about summing two normal distribution in Keras. So, I have 2 normal distributions called epsilon_1 ~ N(0,stdev1²) and epsilon_2 ~ N(0,1 - stdev1²). If I want to sum those 2 normal distribution, is it correct if I only do this epsilon_1 + epsilon_2 in my code?

My Code

stddev1 = 0.25
epsilon_mean = 0
epsilon_std_1 = stddev1
epsilon_std_2 = 1 - stddev1
epsilon1 = K.random_normal(shape=(1, 100),mean=epsilon_mean, stddev=epsilon_std_1)

epsilon2 = K.random_normal(shape=(1, 100),mean=epsilon_mean, stddev=epsilon_std_2)

epsilon = epsilon1 + epsilon2

Upvotes: 0

Views: 139

Answers (1)

Mohammed Khalid
Mohammed Khalid

Reputation: 175

your code is correct, and you can verify by checking the new mean and std of the new random variable, they should be equal to the sum of the originals respectively

Upvotes: 1

Related Questions