[R] Is there a function for this?

[Gabor Grothendieck]

Try this:

> actual <- factor(c("good", "good", "bad", "bad", "good", "good", "bad", "bad"))
> pred <- factor(c("good", "bad", "bad", "bad", "good", "good", "good", "bad"))

> table(actual, pred)
      pred
actual bad good
  bad    3    1
  good   1    3

> prop.table(table(actual, pred), 1)
      pred
actual  bad good
  bad  0.75 0.25
  good 0.25 0.75

> prop.table(table(actual, pred), 2)
      pred
actual  bad good
  bad  0.75 0.25
  good 0.25 0.75

> library(gmodels)
> CrossTable(actual, pred)


   Cell Contents
|-------------------------|
|                       N |
| Chi-square contribution |
|           N / Row Total |
|           N / Col Total |
|         N / Table Total |
|-------------------------|


Total Observations in Table:  8


             | pred
      actual |       bad |      good | Row Total |
-------------|-----------|-----------|-----------|
         bad |         3 |         1 |         4 |
             |     0.500 |     0.500 |           |
             |     0.750 |     0.250 |     0.500 |
             |     0.750 |     0.250 |           |
             |     0.375 |     0.125 |           |
-------------|-----------|-----------|-----------|
        good |         1 |         3 |         4 |
             |     0.500 |     0.500 |           |
             |     0.250 |     0.750 |     0.500 |
             |     0.250 |     0.750 |           |
             |     0.125 |     0.375 |           |
-------------|-----------|-----------|-----------|
Column Total |         4 |         4 |         8 |
             |     0.500 |     0.500 |           |
-------------|-----------|-----------|-----------|





On 1/3/07, Feng Qiu <hustqiufeng at sohu.com> wrote:
> Hi everybody, I'm trying to do a statistic on the error rate of a prediction
> algorithm.
>
> suppose this is the real category
> [good, good, bad, bad, good, good, bad, bad]
> this is the predicted category
> [good, bad, bad, bad, good, good, good, bad]
>
> I'm trying to do a statistic on the error rate for each group("good","bad"):
> what percentage of instances are predicted incorrectly for each group ?
> Of course I can write a loop to do that, but is there a easy way to do that?
>
> Thank you!
>
> Best,
>
> Feng
>