Summarise predictor importance across people
Source:R/stats_prediction_views.R
summarize_importance.RdSummaries use equal weight for each person. Top-feature frequency retains all ties at the requested rank. Pooled and individual means are returned separately rather than averaged together.
Arguments
- x
An idiographic fit or a table returned by
importance().- by
"mean"for mean importance or"frequency"for top-feature counts.- model, scope, subject
Optional filters.
- top_n
Number of highest-ranked features for frequency summaries.
- method, repeats
Passed to
importance()whenxis a fitted object.
Value
A data frame with one row per scope, model, estimator, subgroup and
predictor. Includes n_subjects; frequencies also include count and percent.
Examples
fit <- fit_ml(srl, y = "effort", x = c("efficacy", "monitoring"),
id = "name", model = "ridge", scope = "individual")
summarize_importance(fit, by = "frequency", top_n = 1)
#> scope model estimator subgroup variable n_subjects count percent
#> 1 individual ridge native .none efficacy 36 19 52.77778
#> 2 individual ridge native .none monitoring 36 17 47.22222