Max Kuhn

`classification_cost()` takes probabilities and computes cost from them. Since we are going to be doing more with setting thresholds, we should probably have a version that takes the hard class...

I have the resampling chapter fleshed out but decided to break it in to three parts: 1. Introduction; discussion of general strategies; validation sets; MC-CV; simulation 2. V-fold CV; bootstrapping;...

A reprex from Thomas Dilling over email: ``` r library(tidyverse) library(tidymodels) library(censored) #> Loading required package: survival library(finetune) library(survival) tidymodels_prefer() data(cancer) lung % drop_na() lung % mutate(survival = Surv(time, status...

For some cases, there are too few distinct values of the prediction for the spline to be used.

quantile regression implementation notes

9

Added here to discuss

ensure correct prediction types are used when postprocessing

2

tune currently uses the specified metric sets to determine what types of predictions are being made. For example, for a binary classification model, we might only request "class metrics" (e.g....

If a post process parameter is marked for tuning, we should check what outcomes it changes and then cross-check with the metrics being measured. For example, if someone is optimizing...

> More than one set of outcomes were used when tuning. This should never happen. Review how the outcome is specified in your model. We should list the names of...

reduce the object size by trimming the split elements

2

I had a conversation at conf with someone who mentioned an issue I’ve had. When you have a large data set or a workflow set with many different workflows, the...

probability predictions when using tailor

1

`type = "prob"` returns class predictions. ``` r library(tidymodels) library(tailor) set.seed(816) sim_data % bind_cols(sim_noise(2000, 500)) sim_split 2 class_2 #> 3 class_2 #> 4 class_1 #> 5 class_1 #> 6 class_2...