what are we trying to minimize and why do we consider gradient to be minimized. how are we taking error same as gradient because we are trying to make gradient 0
Linear regression gradient descent
Hey @bajajra,
That’s lots of questions in a single post 
I would recommend you to watch the video again to gain the better understanding of the concept. These basics questions are very well covered in the videos.
Still I try to explain here,
We have made an error function. And we want to find the such parameters that can minimise the error.
Therefore we use gradient descent algorithm to minimise the error function.
Now suppose you are standing on a mountain and you want to reach the valley. What you do is take a step which takes you closer to the bottom line it means you are decreasing height. means you are taking a step in the direction where slope is negative.
The same step we do in gradient descent. You take the derivative ( aka. gradient ) of the error function and update the parameters to the direction in which error function is reducing. Now you reached to a new position where error is lesser than the previous point. In mountain context you reach to position which has a smaller height than earlier position. Again you look for the direction which can take you to even lesser height and you take the step.
Since you are taking steps again and again and every time you are finding the direction in which you should move. Therefore it is an iterative approach. Derivative or gradient tells you the direction where you should take step. And our goal is to reach to point where the error is minimised.very rarely at that position gradient or derivate becomes 0. It means don’t move anything you are already at the most minimum point.
I hope you get the intuition about the algorithm now.
Thanks 
I hope I’ve cleared your doubt. I ask you to please rate your experience here
Your feedback is very important. It helps us improve our platform and hence provide you
the learning experience you deserve.
On the off chance, you still have some questions or not find the answers satisfactory, you may reopen
the doubt.