# How to interpret higher RMSE and $R^2$?

I ran an evaluation on two different regression model variations using the same dataset and I am a bit unsure how to tell which model is better.

The blue model has a worse RMSE of 150 and a better $$R^2$$ score of .107.

I was hoping to see if anyone had any thoughts on how to interpret this in relation to determining which model is more accurate.

Thanks!

• What you’ve described is impossible in math, so there is something about the implementation that is not reporting what you think it is reporting. How are those RMSE and $R^2$ values calculated in the software you’re using?
– Dave
Nov 22, 2022 at 16:35
• Both the r2 are very poor, so saying 0.107 is better than 0.044 is not of much use. Also, the sample count seems different for both. Nov 22, 2022 at 18:55