in the polynomial regression Andrew stated in approx. min 1:45 that the quadratic function would go down, which is not true.
I don’t understand why did we have to choose a cubic function.
Week: 2
Andrew’s explanation mentions why a quadratic function may not always be appropriate: it eventually curves downward, that doesn’t align with the trend you’d expect for housing prices as size increases. This is why a cubic function is introduced (it allows for more flexibility and captures the upward trend while avoiding the unrealistic downturn of a quadratic model).
The choice of a cubic function is an example of feature engineering to better model the data’s behavior.
If you still have questions about this, feel free to share your thoughts!
Thank you for the clarification.
If you suggest that a cubic model serves this purpose better than a quadratic one, I can agree with that. However, stating that a quadratic model will always decrease as x increases is absolutely incorrect.
Best regards,
Abdulrahman
I believe you may be misinterpreting what Andrew is saying.
1 Like
You’re right that a quadratic function doesn’t always decrease as x increases.
Andrew’s point in the video was more about the general shape of quadratic functions, that eventually curve down (or up) for extreme values of x (unsuitable for modeling scenarios like housing prices, which are expected to continuously rise with increasing size).
Hope this helps and clears up any confusion!
I believe you don’t know what a quadratic function means.
Hello Alireza,
I agree with Andrew’s point, and a cubic function seems to be a much better choice in this case. I just wanted to ensure there are no mathematical inaccuracies. Thank you for acknowledging that.
Best regards,
Abdulrahman
1 Like