Adding null model

[EmilHvitfeldt]

in reference to #122.

Still WIP. Most of the skeleton is there but the predict() outputs length of fitting data not new_data.

[topepo]

Please use type = "numeric" for regression and null_model for the user-facing function.

@DavisVaughan will get a laugh out at this coming from me but please lint the R code. Opening brackets should be on the same line as the function definition/if etc.

null_model.default <- function(x = NULL, y, ...) {

}

Can you also enable multivariate numeric regression?

[EmilHvitfeldt]

Can you also enable multivariate numeric regression?

Consider it done!

library(parsnip)

iris_basic <- null_model(mode = "regression") %>% 
  set_engine("parsnip") %>%
  fit(cbind(Sepal.Width) ~ ., data = iris)
  
predict_numeric(iris_basic, iris[1:5,])
#> [1] 3.057333 3.057333 3.057333 3.057333 3.057333

iris_basic <- null_model(mode = "regression") %>% 
  set_engine("parsnip") %>%
  fit(cbind(Sepal.Width, Petal.Width) ~ ., data = iris)
  
predict_numeric(iris_basic, iris[1:5,])
#>   Sepal.Width Petal.Width
#> 1    3.057333    1.199333
#> 2    3.057333    1.199333
#> 3    3.057333    1.199333
#> 4    3.057333    1.199333
#> 5    3.057333    1.199333

^{Created on 2019-01-25 by the reprex package (v0.2.1)}

While we are at it, should we also add quantile and classprob? It doesn't make sense to add conf_int and pred_int.

[topepo]

type = "classprob" would be helpful. I don't think that we have quantiles implemented anywhere so I'd hold off on that.

[EmilHvitfeldt]

Should be there now.

[topepo]

Thanks!

[github-actions]

This pull request has been automatically locked. If you believe you have found a related problem, please file a new issue (with a reprex: https://reprex.tidyverse.org) and link to this issue.