in R I am trying to predict a single value. However, I return all the fitted values when I run the code. I have copied the code directly from my course materials and I get the same result. Am I missing something fundamental?

DF1 <- read.table("linear.txt",header=T)
model <- lm(DF1$Payment~DF1$Claim)
newdata <-data.frame(Claim=4.5)
predict(model,newdata)

my result is

 DF1 <- read.table("linear.txt",header=T)
> DF1
   Claim Payment
1    2.1    2.18
2    2.4    2.06
3    2.5    2.54
4    3.2    2.61
5    3.6    3.67
6    3.8    3.25
7    4.1    4.02
8    4.2    3.71
9    4.5    4.38
10   5.0    4.45
> model <- lm(DF1$Payment~DF1$Claim)
> newdata <-data.frame(Claim=4.5)
> predict(model,newdata)
       1        2        3        4        5        6 
2.016478 2.281170 2.369401 2.987016 3.339938 3.516400 
       7        8        9       10 
3.781092 3.869323 4.134015 4.575168 
Warning message:
'newdata' had 1 row but variables found have 10 rows 
>

when I try different sources of data I get the same situation

my desired out put would be

model$coefficients[1]+model$coefficients[2]*4.5
(Intercept) 
   4.134015 

Thank you very much for your help

in R I am trying to predict a single value. However, I return all the fitted values when I run the code. I have copied the code directly from my course materials and I get the same result. Am I missing something fundamental?

DF1 <- read.table("linear.txt",header=T)
model <- lm(DF1$Payment~DF1$Claim)
newdata <-data.frame(Claim=4.5)
predict(model,newdata)

my result is

 DF1 <- read.table("linear.txt",header=T)
> DF1
   Claim Payment
1    2.1    2.18
2    2.4    2.06
3    2.5    2.54
4    3.2    2.61
5    3.6    3.67
6    3.8    3.25
7    4.1    4.02
8    4.2    3.71
9    4.5    4.38
10   5.0    4.45
> model <- lm(DF1$Payment~DF1$Claim)
> newdata <-data.frame(Claim=4.5)
> predict(model,newdata)
       1        2        3        4        5        6 
2.016478 2.281170 2.369401 2.987016 3.339938 3.516400 
       7        8        9       10 
3.781092 3.869323 4.134015 4.575168 
Warning message:
'newdata' had 1 row but variables found have 10 rows 
>

when I try different sources of data I get the same situation

my desired out put would be

model$coefficients[1]+model$coefficients[2]*4.5
(Intercept) 
   4.134015 

Thank you very much for your help

• r

1 Answer 1

Don't use $ in your formula when fitting the model, use the data argument instead:

lm(Payment ~ Claim, data = DF1)

if you use DF1$Payment etc., R doesn't know that it should substitute the variable in your new (prediction) data set ...

If you have indeed copied exactly what was in the course materials, you should let your instructor know that not specifying data/using $ in formulas is generally bad practice, for this reason ... (even if you don't want to use predict afterwards, using the data argument makes the model call more readable)