SOLUTION: i have for a long time been puzzled by the question as to whether there is a real solution for determining in advance which equation system (among others, linear quadratic etc...)

Algebra ->  Linear-equations -> SOLUTION: i have for a long time been puzzled by the question as to whether there is a real solution for determining in advance which equation system (among others, linear quadratic etc...)       Log On


   

Question 1108180: i have for a long time been puzzled by the question as to whether there is a real solution for determining in advance which equation system (among others, linear quadratic etc...) best fit the scenario of depicting a strong association between dependent, experimented , controlled and independent variables?
Answer by Theo(13342) About Me  (Show Source):
You can put this solution on YOUR website!
i believe the r^2 will tell you that.

r^2 tells you how well the model fits your data.

if r^2 is high, then you have a good fit.

if r^s is not so high, then you don't have a good fit.

if you're not sure, you would model using the different types of modeling techniques and then look at the r^2 for each.

the better fitting model would be the one with the higher r^2.

i'm not an expert, but that's what i would do if i was in doubt as to which model gave me the best fit for the data.

here's a more complex method taken from the web which i don't understand fully.

http://statisticsbyjim.com/regression/model-specification-variable-selection/

here's another one from the web.

https://www.theanalysisfactor.com/assessing-the-fit-of-regression-models/

lots more on the web.
i did a search for "how to determine the type of regression model to use"
and some of the options are the ones shown above.

it appears that r^2 is a good test, but it's not quite so simple as it might seem at first and other types of test might be required as well.