predict_terms: predict_terms

In hydrorecipes: Hydrogeology Steps for the 'recipes' Package

predict_terms

Description

Predict the contribution for each step.

Usage

predict_terms(fit, rec, data, ...)

## S3 method for class 'lm'
predict_terms(fit, rec, data, ...)

## S3 method for class 'cv.glmnet'
predict_terms(fit, rec, data, ...)

## S3 method for class 'numeric'
predict_terms(fit, rec, data, ...)

Arguments

fit	A model object that has a coefficients method (e.g. lm)
rec	A prepped recipe
data	A data.frame with feature columns
...	Currently not used

Value

A data.frame of predicted values for each step (component contribution).

Examples

data(transducer)
transducer$datetime_num <- as.numeric(transducer$datetime)

rec_toll_rasmussen <- recipe(wl ~ baro + et + datetime_num, transducer) |>
   step_lead_lag(baro, lag = log_lags(100, 86400 * 2 / 120)) |>
   step_ns(datetime_num, deg_free = 10) |>
   prep()

input_toll_rasmussen <- rec_toll_rasmussen |> bake(new_data = NULL)

fit_toll_rasmussen <- lm(wl ~ ., input_toll_rasmussen)
pred <- predict_terms(fit_toll_rasmussen,
                      rec_toll_rasmussen,
                      input_toll_rasmussen)

hydrorecipes documentation built on June 27, 2022, 9:06 a.m.