This is what am trying to do:
| Time | value |
|---|---|
| 00:00 | 12 |
| 00:01 | 6 |
| 00:02 | 8 |
| 00:03 | 12 |
| 00:04 | 11 |
| 00:05 | 6 |
| 00:06 | 13 |
| 00:07 | 12 |
| 00:08 | 12 |
and i want sample variance and result like:
| Time | value | samp_vars |
|---|---|---|
| 00:00 | 12 | NA |
| 00:01 | 6 | NA |
| 00:02 | 8 | 24.1666 |
| 00:03 | 12 | 12.8 |
| 00:04 | 11 | 6.8 |
| 00:05 | 6 | 3.0667 |
| 00:06 | 13 | 3.7 |
| 00:07 | 12 | 3.5833 |
| 00:08 | 12 | 0.333 |
I tried doing this:
df$samp_vars <- rollapply(df$value, list(c(-2,0,3)), var, fill = NA)
but my variance sample was not correct. I am not an expert in R so i will appreciate the help. Thanks!
CodePudding user response：
I make the expected values to be NA, NA, 8.166, 9.466, 7.466, 6.400, NA, NA, NA — not as you have in the question. For example, at 00:02
var(c(12,6,8,12,11,6))
To get those values, simply alter your existing code to use a sequence of the right integers using : — at the moment you are getting the variance at the current row, the row two steps back and the row three steps forward, not all in between.
zoo::rollapply(data$value, list(-2:3), var, fill = NA)
The -2:3 is making a vector of the integers c(-2, -1, 0, 1, 2, 3) for you.