0

I have two regression model one with Correlation Coefficient of 87 and RMSE of 0.045 and the other has Correlation Coefficient of 93 and RMSE of 0.055. Which of them is better model for prediction of future cases? OR which of Correlation Coefficient or RMSE is better for regression model evaluation? Thanks.

merv
  • 67,214
  • 13
  • 180
  • 245
nasaj nasaj
  • 9
  • 1
  • 2
  • What does RMSE stand for? Are you sure Stackoverflow is the right place for this question? You're expecting an answer from someone who both knows what is RMSE and how to evaluate a regression model, maybe in statictics or math you will get more answers. Also, are you sure google couldn't help you? – Vince Sep 26 '12 at 12:51

1 Answers1

2

Two variable Correlation Coefficient and RMSe don't have relation with each other. Correlation Coefficient show that weight of relation between these item. when Correlation Coefficient is high so it's good. RMSE show the error at the result depend on input when RMSE is low so it's ideal. according to your business you can select. on my business i'll select second item Correlation Coefficient of 93 and RMSE of 0.055 because i more sure that items have a relation with each other and RMSE = 0.055 is very small.

Ahmed Mongy
  • 76
  • 5