Advance Gradient Descent

In Simple BGD (Batch Gradient Descent) the parameters are updated by some of all the square error of data set. However, BGD convergence much accurately but when there are large data set then the convergence will take lot of time and memory. Therefore, to overcome these problems, following evolution were made on optimizing parameters. Stochastic … Continue reading Advance Gradient Descent

Gradient Descent and Cost Functions

When talking about linear regression(only one variable) then the hypothesis is ℎ𝜃 (𝑥) = 𝜃0 + 𝜃1𝑥 where hypothesis is best fitting line for the data. In this equation, guessing theta values that creates the best fit line for our linear regression model is important. Therefore to determine the theta values in the equation, there … Continue reading Gradient Descent and Cost Functions