Computing accuracy score on Stata

[Dominique Bourget]

Hello,

I am trying to compare two Stata variables "yBestPred" and "compliance" and try to come up with the accuracy score.
To compute the accuracy score, I need to divide the number of observations that shares the same value (either 0 or 1) for both yBestPred and compliance by the total number of observation.

Let's say that my variables look something like below:

>> list yBestPred

1.  1
2.  1
3.  1
4.  0
5.  1

>> list compliance

1.  1
2.  0
3.  1
4.  1
5.  1

How can I do this in Stata? I know how to do this on Python, but just don't know how to do this on Stata.

Thanks,

[daniel klein]

There are many ways to do this. One is

clear
input yBestPred compliance
1 1
1 0
1 1
0 1
1 1
end

// accuracy
count if yBestPred==compliance
display r(N)/c(N)

Note that this approach is very basic, not very flexible, and not robust at all. For example, selecting a subset of observations would require more code; missing values are likely to mess up the result, etc.

Best
Daniel

[daniel klein]

Discussion continued here.