Search notes:
R function: predict
# X11()
d <- data.frame (
t <- c(25, 22, 18, 19, 21, 26, 26, 27, 24, 22),
w <- c( 4, 3, 0, 6, 5, 3, 4, 1, 3, 1)
)
At = 3
Bt = 4
Et = rnorm(length(t), mean = 0, sd = 3.2)
Aw = 5
Bw = -2
Ew = rnorm(length(t), mean = 0, sd = 0.9)
d$o <- At + Bt*t + Et +
Aw + Bw*w + Ew
model <- lm(d$o ~ d$t * d$w)
prediction <- predict(model)
data.frame(
p = prediction,
o = d$o
)
# p o
# 1 101.92822 101.02235
# 2 90.43924 93.38205
# 3 79.93667 80.02431
# 4 71.49645 70.13953
# 5 82.57825 84.57244
# 6 107.48234 106.16260
# 7 106.31569 107.85726
# 8 113.82305 115.09546
# 9 98.96079 95.83189
# 10 93.78607 92.65888
cat ("\n\n")
predict(model, data.frame(t=20, w=5))
# Warning message:
# 'newdata' had 1 row but variables found have 10 rows
# TODO http://stackoverflow.com/questions/31200748&
# plot(prediction, which=c(2))
# z <- locator(1)