Hi Abul, Good question!

The 1/2 is simply for convenience. If you take the derivative of X², then it becomes 2X. 1/2 is used to remove that 2.

1/m is to calculate the “mean” squared error. Whether m (sample size) is small or big, we can make a fair comparison this way.