incorrect calculating (R-squared) in R(wrong value)

Question

Closed. This question is not reproducible or was caused by typos. It is not currently accepting answers. question from:https://stackoverflow.com/questions/65920932/incorrect-calculating-r-squared-in-rwrong-value

Answer 1

It is likely that x is not the predictions but is a predictor that goes into a linear regression. Perform the regression, fm, in which case the predicted values are fitted(fm) and then get the R squared from summary or get it directly as shown in the alternatives.

fm <- lm(yield ~ x, mydat)
summary(fm)$r.squared
# [1] 0.02508245

# same
cor(mydat$yield, fitted(fm))^2
# [1] 0.02508245

# same
with(mydat, cor(yield, x)^2)
# [1] 0.02508245

# same
tss <- with(mydat, sum((yield - mean(yield))^2))
rss <- deviance(fm)
1 - rss/tss
# [1] 0.02508245

# same
tss <- with(mydat, sum((yield - mean(yield))^2))
rss <- sum(resid(fm)^2)
1 - rss/tss
# [1] 0.02508245


plot(yield ~ x, mydat)
abline(fm)