Error with h2o.explain

[pedro-w]

H2O version, Operating System and Environment

• H2O cluster version: 3.44.0.3
• OpenJDK Runtime Environment Temurin-21.0.4+7 (build 21.0.4+7-LTS)
• R Version: R version 4.4.1 (2024-06-14 ucrt)
• R studio 2024.12.0 Build 467
• Windows 11 / x64

Actual behavior
Attempting to use h2o.explain on a h20.adaBoost model gives an error

Error in if (is.na(i)) NULL else x[, i] : argument is of length zero

Expected behavior

Error should not occur (tried with h2o.naiveBayes model and it worked successfully)

Steps to reproduce

• Source minimal R script (below)
• Error appears on console output

library(h2o)
h2o.init()
# Import the prostate dataset into H2O:
prostate <- h2o.importFile("https://s3.amazonaws.com/h2o-public-test-data/smalldata/prostate/prostate.csv")
predictors <- c("AGE","RACE","DPROS","DCAPS","PSA","VOL","GLEASON")
response <- "CAPSULE"
prostate[response] <- as.factor(prostate[response])
# Build and train the model:
adaboost_model <- h2o.adaBoost(nlearners=50,
                               learn_rate = 0.5,
                               weak_learner = "DRF",
                               x = predictors,
                               y = response,
                               training_frame = prostate)

# Generate predictions:
h2o.predict(adaboost_model, prostate)
# Explain model
h2o.explain(adaboost_model, prostate)

[tomasfryda]

Thank you for reporting this issue. AdaBoost is currently does not support explain.

[pedro-w]

Thank you for the swift response.