Question

I want to calculate the variance in a dataset collected at a given time (t) for which we know the frequency of the conditional and decision attributes.

Every conditional attribute can take 3 possible values while the outcome consists of a two-classe attribute, as can be seen in the picture given below. What I need is to calculate the variance of this given data snapshot.

I would also be interested in sample code or algorithm.

!

Answer 1

I think you would be better off asking this on the “statsOverflow” page 🙂

The question above is a bit unclear to me, but in discrete cases you would calculate the variance with the following algorithm:

data_type getVariance(data_type[] array){
    data_type mean = getMean(array);
    data_type var = 0;
    for(int i =0; i < array.length; i++){
        var += getProbabilityOf(array[i])*(array[i]-mean)*(array[i]-mean);
    }
    return var;
}
data_type getMean(data_type[] array){
    data_type mean = 0;
    for(int i =0; i < array.length; i++){
        mean += getProbabilityOf(array[i], array)*array[i];
    }
    return mean;
}
double getProbabilityOf(data_type data, data_type[] array){
    return [number of appearances of data in array / size of array];
}