In project 13 on polynomial regression, i observed that upon changing the no. of co-efficient value, there has been no change in the prediction plot. I varied the value from 2 to 28 which produce no changes, but 29 and 30 causes over-fitting.

Are the changes not observed for just this data-set in the project? and

How does the value effect the accuracy of the predictions?

Thank you

@suhailc0088 changing the number of coefficient is actually changing the order of polynomial, i don’t know happened in your case but changing parameters actually effects your predicted data.

We know that lesser orders might lead to under-fitting while the higher might lead to over-fitting, so that’s why you have to choose the number which bests fit the data. Unfortunately you can’t get accuracy score here which tells you accuracy of your model over different values that’s why you have to do things manually which means choose parameters which best fits predicted line with your actual line…I HOPE IT HELPED…!!!