Hi, When running the eSHAP_plot() function, I encounter an error in the mutate() step where forcats::fct_reorder(feature, mean_phi) fails. The error message indicates that lvls_reorder() expects idx to contain one integer for each level of f, but this condition is not met.
rlang::last_trace()
<error/dplyr:::mutate_error>
Error in `mutate()`:
ℹ In argument: `feature = forcats::fct_reorder(feature,
mean_phi)`.
Caused by error in `lvls_reorder()`:
! `idx` must contain one integer for each level of `f`
---
Backtrace:
▆
1. ├─explainer::eSHAP_plot(...)
2. │ └─... %>% ggplot(aes(x = feature, y = Phi, color = f_val))
3. ├─ggplot2::ggplot(., aes(x = feature, y = Phi, color = f_val))
4. ├─dplyr::mutate(., feature = forcats::fct_reorder(feature, mean_phi))
5. ├─dplyr:::mutate.data.frame(., feature = forcats::fct_reorder(feature, mean_phi))
6. │ └─dplyr:::mutate_cols(.data, dplyr_quosures(...), by)
7. │ ├─base::withCallingHandlers(...)
8. │ └─dplyr:::mutate_col(dots[[i]], data, mask, new_columns)
9. │ └─mask$eval_all_mutate(quo)
10. │ └─dplyr (local) eval()
11. └─forcats::fct_reorder(feature, mean_phi)
12. └─forcats::lvls_reorder(f, order(summary, decreasing = .desc))mydata <- read.csv('train.csv')
library(tidyverse)
mydata <- mydata %>%
mutate(across(where(is.character), as.factor)) %>%
mutate(across(where(is.numeric), as.numeric)) %>%
mutate(Disease = as.factor(Disease))
# Create a classification task
maintask <- mlr3::TaskClassif$new(
id = "my_classification_task",
backend = mydata,
target = 'Disease',
positive = '1'
)
# Split the dataset for training
splits <- mlr3::partition(maintask)
# Create a ranger learner for classification
mylrn <- mlr3::lrn("classif.ranger", predict_type = "prob")
# Train the learner on the training set
mylrn$train(maintask, splits$train)
# Generate SHAP values and plot
SHAP_output <- eSHAP_plot(
task = maintask,
trained_model = mylrn,
splits = splits,
sample.size = 30,
seed = seed
)
Hi, When running the
eSHAP_plot()function, I encounter an error in the mutate() step whereforcats::fct_reorder(feature, mean_phi) fails. The error message indicates thatlvls_reorder()expects idx to contain one integer for each level of f, but this condition is not met.data:
train.csv
error trace