@lokesh_jindal wrote:
Hi, I have seen people trying to run polynomial regression of second order (quadratic regression) with a single variable by using x and x^2 as the independent variables in excel itself. My doubt is that, wont the model suffer due to strong multicollinearity between its independent variables. There is also a counter argument saying that x and x^2 do not have linear relationship and hence model wont suffer. But when the data has only positive value of x (say an independent variable like age) then it does have strong correlation with X^2. In other words, the variable is not centered. Here is a you tube video that i found out wherein the user creates such a model in excel. https://www.youtube.com/watch?v=kFasmbrDG5I&t=8s
Is the approach shown in the video wrong? If yes then how to find the hidden mistake and also how does the model still fits the input data properly.
Thanks in advance
Posts: 1
Participants: 1