After adding the x^3 feature to the training set and plotting it, the text says, “Gradient descent has emphasized the data that is the best fit to the x^2 data by increasing the w_1 term relative to the others.” I believe this should be, '…increasing the w_2 term relative to the others."? This is the model at this stage, “0.08𝑥+0.54𝑥2+0.03𝑥3+0.0106”.

Thank-you.

Hi @chantrym

You are correct, based on the image above (taken from course slides), the weight indices start from 1 and 0.54 is the coefficient w_2 associated with x^2. So, it must be w_2 instead of w_1.

Thanks for reporting the issue!

Would you mind taking a look at this? @balaji.ambresh

Thanks for considering me, @Alireza_Saei

Unfortunately, I didn’t take up the new version of MLS. Adding @rmwkwok and @TMosh since it’s unclear to me on how the weights are organized.

Hello, @Alireza_Saei, in the " Selecting Features" section, in the first paragraph, the equation at the end states that they are associating w_1 with the squared term.

Hi @rmwkwok

Thank you for the clarification. Based on the image shared from the course slides, I thought the weight indices start from 1 (or we assign w_0 to the bias term).

Thanks again!

My apologies. Thank-you for clarifying.