stratified_sampling

stratified_sampling is a child class of Sampler class. It randomly samples alternatives from choice population according to their stratum setting. Its run method accepts the following arguments:

dataset1

- an instance of Dataset, to be used as agent set.

dataset2

- an instance of Dataset, to be used as choice set.

index1

- indices of dataset1 for whom alternatives are sampled. If it is not given, all elements of dataset1 are used.

index2

- indices of dataset2 from which alternatives are sampled. If it is not given, all elements of dataset2 are used.

stratum

- an array indicates the stratum id for elements of dataset2; it has to be either of None, or of the same size as index2 or dataset2. If it's not given, all elements are treated as in 1 stratum.

weight

- like in weighted_sampling, weight is an array used as weight for elements of dataset2 in unequal probability sampling; it has to be either of None, or of the same size as index2 or dataset2. If it is not given, sampling is proceeded with equal probability.

sample_size

- number of alternatives sampled from one stratum; default value is 1.

sample_size_from_chosen_stratum

- number of alternatives sampled from agent's chosen stratum. If it's None, it's equal to value specified by sample_size or sample|_rate.

sample_rate

- calculate number of alternatives sampled from one stratum by multiplying this rate with number of observations in this stratum. If both sample_rate and sample_size are specified, use sample_rate.

include_chosen_choice

- whether agents' chosen choice will be included in the return results. If it's true, the chosen choices are in the first column of the return results.

resources

- an instance of Resources that can be used to pass any of the above arguments to run method.