What would be considered an acceptable value for Std. Error of the Estimate of a regression?
Hi dear friend This is strange questions that have no answers. In a study, the number 1 for the Std. Error in the regression estimation may be too high, and in the other research, the number 1000 is also very good. Everything depends on your numbers and data. In fact, the Std. Error is not a valid statistic for examining the regression model. Instead, the value of R2 and the probability values of the ANOVA table and the regression coefficients of our estimation source will be.
Dear Abolfazl,
Actually I am working with non-linear regression more precisely to find out best fit curve for carbon emissions with different socio-economic factors (GDP, Population, Savings etc.). I have read a post where it says R square is not a valid test for non-linear regression curve fitting estimates but std. error is http://statisticsbyjim.com/regression/r-squared-invalid-nonlinear-regression/. Moreover all results are significant at 0.00 under ANOVA.
I have sent you the formula for the R square. I do not think so. R square measures the relationship between estimation and reality and is always useful. However, since your regression coefficients are meaningful, you get a good model. In this regard, Minitab software is very good.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression