I have two equally long matching vectors of time series data: Price (x) and hour (h). Hour goes from 0-23. My hour variable is my dummy variable (or factor/level variable I guess it is called in R).
Right now i've defined 24 different dummy variables, and for each hour I type my dummy variable. So for example generating 24 plots to look at or calculate 24 means etc I would type: plot.ts(hour1) # and so on for all 24.
I would like to do this for all 24 variables as easily as possible? So I can run a lot of different calculations. For example, how could I just compute the mean for all 24 dummy variables without making 24 lines of code, changing each dummy variable?
EDIT: Sorry, thought it was clear with the two vectors. Example:
1. Price Hour
2. 8 0
3. 12 1
4. 14 2
5. 16 3
6. 18 4
7. 20 5
8. 22 6
9. 24 7
10. 26 8
11. 28 9
12. 24 10
13. 26 11
14. 23 12
15. 23 13
16. 23 14
17. 14 15
18. 19 16
19. 25 17
20. 26 18
21. 28 19
22. 30 20
23. 33 21
24. 24 22
25. 10 23
26. 14 0
27. 12 1
28. 13 2
29. x ect.
It is not clear how your data are stored since you don't give a reproducible example. I assume you have separate variables for each hour1.
Generally, It is better to put your hourxx variable in a list to perform calculations.
For example, this will compute mean for all hours:
lapply(lapply(ls(pattern='hour.*'),get),mean)
EDIT after OP clarification:
You shuld create a new variable to distinguish between Hours intervals. Something like :
dat <- data.frame(Price=rnorm(24*5),Hour=rep(0:23,5))
dat$id <- cumsum(c(0,diff(dat$Hour)==-23))
Then using ply package for example , you can compute mean by id:
library(plyr)
ddply(dat,.(id),summarise,mPrice=mean(Price))
id mPrice
1 0 0.2999602
2 1 -0.2201148
3 2 0.2400192
4 3 -0.2087594
5 4 0.1666915