Need help with model choosing

[gx2144]

Hi, I have trouble with selecting proper model to fit the data. I first apply the simple linear regression with one independent variable and one dependent variable. However, it fits bad. I would like to now what model is appropriate for data distribution like this? Or what can I do to improve the fit?

[PaigeMiller · Accepted solution]

This variable cannot be used to predict the response (other than poorly).

You need to try a different variable.

[Ksharp]

Your data look like a X correspond to many Y . Could try PROC QUANREG (quantitle regression). @Rick_SAS  wrote a couple of blogs about it .