LIME: sampling for local exploration by permuting all columns.
permutation_neighbourhood(data, explained_instance, size, fixed_variables, seed = NULL)
data | Data frame from which observations will be generated. |
---|---|
explained_instance | A row in an original data frame (as a data.frame). |
size | Number of observations to be generated. |
fixed_variables | Names of column which will not be changed while sampling. |
seed | Seed to set before sampling. If NULL, results will not be reproducible. |
data frame