Skip to content

Update xgboost learner to latest CRAN version #31

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 26 commits into from
Nov 20, 2019
Merged

Update xgboost learner to latest CRAN version #31

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
26 commits
Select commit Hold shift + click to select a range
4ec1e3c
upd xgboost learner to latest cran version
pat-s Sep 7, 2019
1f98fbb
correct constructor example and typo
pat-s Sep 7, 2019
7b0197f
add param deps for xgboost
pat-s Sep 8, 2019
e29fff2
Increment version number
pat-s Sep 8, 2019
573432f
add patrick as aut
pat-s Sep 8, 2019
2f24085
upd news
pat-s Sep 8, 2019
4a33f50
Merge branch 'master' into xgboost
pat-s Sep 8, 2019
e300eaf
fix regr.xgboost param
pat-s Sep 8, 2019
8ec4baa
Merge branch 'xgboost' of github.com:mlr-org/mlr3learners into xgboost
pat-s Sep 8, 2019
b83b132
error on warning on Travis
pat-s Sep 9, 2019
3fac921
Add colsample_bynode parameter
be-marc Oct 15, 2019
11e8823
Fix tree_method
be-marc Oct 15, 2019
babded3
Add sketch_eps
be-marc Oct 15, 2019
5892d71
Add scale_pos_weight
be-marc Oct 15, 2019
84ae34f
Fix parameter dependencies
be-marc Oct 15, 2019
723833c
Add lambda, lambda_bias and alpha dependency on gblinear
be-marc Oct 15, 2019
2adb6ba
Add updater parameter
be-marc Oct 15, 2019
6910a10
Add missing tags
be-marc Oct 15, 2019
0e67c57
Add refresh_leaf
be-marc Oct 15, 2019
fff67e5
Add feature_selector and top_k
be-marc Oct 15, 2019
9ef5c84
Add changes to LearnerRegrXgboost
be-marc Oct 16, 2019
3144f4e
Merge branch 'master' into xgboost
be-marc Oct 16, 2019
09661bf
Add predictor parameter
be-marc Oct 16, 2019
017c339
Update NEWS.md
pat-s Nov 11, 2019
daddc2d
Merge branch 'master' into xgboost
be-marc Nov 12, 2019
e4f44e2
Merge branch 'xgboost' of github.com:mlr-org/mlr3learners into xgboost
be-marc Nov 12, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions NEWS.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,9 @@
* Improved documentation.
* Added references.

* add new parameters of xgboost version 0.90.2
* add parameter dependencies for xgboost

# mlr3learners 0.1.2

* Maintenance release.
Expand Down
58 changes: 43 additions & 15 deletions R/LearnerClassifXgboost.R
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
#' @title Extreme Gradiant Boosting Classification Learner
#' @title Extreme Gradient Boosting Classification Learner
#'
#' @usage NULL
#' @aliases mlr_learners_classif.xgboost
Expand Down Expand Up @@ -30,11 +30,12 @@ LearnerClassifXgboost = R6Class("LearnerClassifXgboost", inherit = LearnerClassi
ParamUty$new("watchlist", default = NULL, tags = "train"),
ParamDbl$new("eta", default = 0.3, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("gamma", default = 0, lower = 0, tags = "train"),
ParamInt$new("max_depth", default = 6L, lower = 1L, tags = "train"),
ParamInt$new("max_depth", default = 6L, lower = 0L, tags = "train"),
ParamDbl$new("min_child_weight", default = 1, lower = 0, tags = "train"),
ParamDbl$new("subsample", default = 1, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("colsample_bytree", default = 1, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("colsample_bylevel", default = 1, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("colsample_bynode", default = 1, lower = 0, upper = 1, tags = "train"),
ParamInt$new("num_parallel_tree", default = 1L, lower = 1L, tags = "train"),
ParamDbl$new("lambda", default = 1, lower = 0, tags = "train"),
ParamDbl$new("lambda_bias", default = 0, lower = 0, tags = "train"),
Expand All @@ -45,26 +46,53 @@ LearnerClassifXgboost = R6Class("LearnerClassifXgboost", inherit = LearnerClassi
ParamDbl$new("max_delta_step", lower = 0, default = 0, tags = "train"),
ParamDbl$new("missing", default = NA, tags = c("train", "predict"), special_vals = list(NA, NA_real_, NULL)),
ParamInt$new("monotone_constraints", default = 0L, lower = -1L, upper = 1L, tags = "train"),
ParamDbl$new("tweedie_variance_power", lower = 1, upper = 2, default = 1.5, tags = "train"), # , requires = quote(objective == "reg:tweedie")
ParamDbl$new("tweedie_variance_power", lower = 1, upper = 2, default = 1.5, tags = "train"),
ParamInt$new("nthread", lower = 1L, tags = "train"),
ParamInt$new("nrounds", lower = 1L, tags = "train"),
ParamUty$new("feval", default = NULL, tags = "train"),
ParamInt$new("verbose", default = 1L, lower = 0L, upper = 2L, tags = "train"),
ParamInt$new("print_every_n", default = 1L, lower = 1L, tags = "train"), # , requires = quote(verbose == 1L)
ParamInt$new("print_every_n", default = 1L, lower = 1L, tags = "train"),
ParamInt$new("early_stopping_rounds", default = NULL, lower = 1L, special_vals = list(NULL), tags = "train"),
ParamLgl$new("maximize", default = NULL, special_vals = list(NULL), tags = "train"),
ParamFct$new("sample_type", default = "uniform", levels = c("uniform", "weighted"), tags = "train"), # , requires = quote(booster == "dart")
ParamFct$new("normalize_type", default = "tree", levels = c("tree", "forest"), tags = "train"), # , requires = quote(booster == "dart")
ParamDbl$new("rate_drop", default = 0, lower = 0, upper = 1, tags = "train"), # , requires = quote(booster == "dart")
ParamDbl$new("skip_drop", default = 0, lower = 0, upper = 1, tags = "train"), # , requires = quote(booster == "dart")
# TODO: uncomment the following after the next CRAN update, and set max_depth's lower = 0L
# ParamLgl$new("one_drop", default = FALSE, requires = quote(booster == "dart")),
# ParamFct$new("tree_method", default = "exact", levels = c("exact", "hist"), requires = quote(booster != "gblinear")),
# ParamFct$new("grow_policy", default = "depthwise", levels = c("depthwise", "lossguide"), requires = quote(tree_method == "hist")),
# ParamInt$new("max_leaves", default = 0L, lower = 0L, requires = quote(grow_policy == "lossguide")),
# ParamInt$new("max_bin", default = 256L, lower = 2L, requires = quote(tree_method == "hist")),
ParamUty$new("callbacks", default = list(), tags = "train")
ParamFct$new("sample_type", default = "uniform", levels = c("uniform", "weighted"), tags = "train"),
ParamFct$new("normalize_type", default = "tree", levels = c("tree", "forest"), tags = "train"),
ParamDbl$new("rate_drop", default = 0, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("skip_drop", default = 0, lower = 0, upper = 1, tags = "train"),
ParamLgl$new("one_drop", default = FALSE, tags = "train"),
ParamFct$new("tree_method", default = "auto", levels = c("auto", "exact", "approx", "hist", "gpu_hist"), tags = "train"),
ParamFct$new("grow_policy", default = "depthwise", levels = c("depthwise", "lossguide"), tags = "train"),
ParamInt$new("max_leaves", default = 0L, lower = 0L, tags = "train"),
ParamInt$new("max_bin", default = 256L, lower = 2L, tags = "train"),
ParamUty$new("callbacks", default = list(), tags = "train"),
ParamDbl$new("sketch_eps", default = 0.03, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("scale_pos_weight", default = 1, tags = "train"),
ParamUty$new("updater", tags = "train"), # Default depends on the selected booster
ParamLgl$new("refresh_leaf", default = TRUE, tags = "train"),
ParamFct$new("feature_selector", default = "cyclic", levels = c("cyclic", "shuffle", "random", "greedy", "thrifty"), tags = "train"),
ParamInt$new("top_k", default = 0, lower = 0, tags = "train"),
ParamFct$new("predictor", default = "cpu_predictor", levels = c("cpu_predictor", "gpu_predictor"), tags = "train")
))
# param deps
ps$add_dep("tweedie_variance_power", "objective", CondEqual$new("reg:tweedie"))
ps$add_dep("print_every_n", "verbose", CondEqual$new(1L))
ps$add_dep("sample_type", "booster", CondEqual$new("dart"))
ps$add_dep("normalize_type", "booster", CondEqual$new("dart"))
ps$add_dep("rate_drop", "booster", CondEqual$new("dart"))
ps$add_dep("skip_drop", "booster", CondEqual$new("dart"))
ps$add_dep("one_drop", "booster", CondEqual$new("dart"))
ps$add_dep("tree_method", "booster", CondAnyOf$new(c("gbtree", "dart")))
ps$add_dep("grow_policy", "tree_method", CondEqual$new("hist"))
ps$add_dep("max_leaves", "grow_policy", CondEqual$new("lossguide"))
ps$add_dep("max_bin", "tree_method", CondEqual$new("hist"))
ps$add_dep("sketch_eps", "tree_method", CondEqual$new("approx"))
ps$add_dep("lambda", "booster", CondEqual$new("gblinear"))
ps$add_dep("lambda_bias", "booster", CondEqual$new("gblinear"))
ps$add_dep("alpha", "booster", CondEqual$new("gblinear"))
ps$add_dep("feature_selector", "booster", CondEqual$new("gblinear"))
ps$add_dep("top_k", "booster", CondEqual$new("gblinear"))
ps$add_dep("top_k", "feature_selector", CondAnyOf$new(c("greedy", "thrifty")))

# custom defaults
ps$values = list(nrounds = 1L, verbose = 0L)

super$initialize(
Expand Down
56 changes: 42 additions & 14 deletions R/LearnerRegrXgboost.R
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,11 +30,12 @@ LearnerRegrXgboost = R6Class("LearnerRegrXgboost", inherit = LearnerRegr,
ParamUty$new("watchlist", default = NULL, tags = "train"),
ParamDbl$new("eta", default = 0.3, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("gamma", default = 0, lower = 0, tags = "train"),
ParamInt$new("max_depth", default = 6L, lower = 1L, tags = "train"),
ParamInt$new("max_depth", default = 6L, lower = 0L, tags = "train"),
ParamDbl$new("min_child_weight", default = 1, lower = 0, tags = "train"),
ParamDbl$new("subsample", default = 1, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("colsample_bytree", default = 1, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("colsample_bylevel", default = 1, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("colsample_bynode", default = 1, lower = 0, upper = 1, tags = "train"),
ParamInt$new("num_parallel_tree", default = 1L, lower = 1L, tags = "train"),
ParamDbl$new("lambda", default = 1, lower = 0, tags = "train"),
ParamDbl$new("lambda_bias", default = 0, lower = 0, tags = "train"),
Expand All @@ -45,26 +46,53 @@ LearnerRegrXgboost = R6Class("LearnerRegrXgboost", inherit = LearnerRegr,
ParamDbl$new("max_delta_step", lower = 0, default = 0, tags = "train"),
ParamDbl$new("missing", default = NA, tags = c("train", "predict"), special_vals = list(NA, NA_real_, NULL)),
ParamInt$new("monotone_constraints", default = 0L, lower = -1L, upper = 1L, tags = "train"),
ParamDbl$new("tweedie_variance_power", lower = 1, upper = 2, default = 1.5, tags = "train"), # , requires = quote(objective == "reg:tweedie")
ParamDbl$new("tweedie_variance_power", lower = 1, upper = 2, default = 1.5, tags = "train"),
ParamInt$new("nthread", lower = 1L, tags = "train"),
ParamInt$new("nrounds", lower = 1L, tags = "train"),
ParamUty$new("feval", default = NULL, tags = "train"),
ParamInt$new("verbose", default = 1L, lower = 0L, upper = 2L, tags = "train"),
ParamInt$new("print_every_n", default = 1L, lower = 1L, tags = "train"), # , requires = quote(verbose == 1L)
ParamInt$new("print_every_n", default = 1L, lower = 1L, tags = "train"),
ParamInt$new("early_stopping_rounds", default = NULL, lower = 1L, special_vals = list(NULL), tags = "train"),
ParamLgl$new("maximize", default = NULL, special_vals = list(NULL), tags = "train"),
ParamFct$new("sample_type", default = "uniform", levels = c("uniform", "weighted"), tags = "train"), # , requires = quote(booster == "dart")
ParamFct$new("normalize_type", default = "tree", levels = c("tree", "forest"), tags = "train"), # , requires = quote(booster == "dart")
ParamDbl$new("rate_drop", default = 0, lower = 0, upper = 1, tags = "train"), # , requires = quote(booster == "dart")
ParamDbl$new("skip_drop", default = 0, lower = 0, upper = 1, tags = "train"), # , requires = quote(booster == "dart")
# TODO: uncomment the following after the next CRAN update, and set max_depth's lower = 0L
# ParamLgl$new("one_drop", default = FALSE, requires = quote(booster == "dart")),
# ParamFct$new("tree_method", default = "exact", levels = c("exact", "hist"), requires = quote(booster != "gblinear")),
# ParamFct$new("grow_policy", default = "depthwise", levels = c("depthwise", "lossguide"), requires = quote(tree_method == "hist")),
# ParamInt$new("max_leaves", default = 0L, lower = 0L, requires = quote(grow_policy == "lossguide")),
# ParamInt$new("max_bin", default = 256L, lower = 2L, requires = quote(tree_method == "hist")),
ParamUty$new("callbacks", default = list(), tags = "train")
ParamFct$new("sample_type", default = "uniform", levels = c("uniform", "weighted"), tags = "train"),
ParamFct$new("normalize_type", default = "tree", levels = c("tree", "forest"), tags = "train"),
ParamDbl$new("rate_drop", default = 0, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("skip_drop", default = 0, lower = 0, upper = 1, tags = "train"),
ParamLgl$new("one_drop", default = FALSE, tags = "train"),
ParamFct$new("tree_method", default = "auto", levels = c("auto", "exact", "approx", "hist", "gpu_hist"), tags = "train"),
ParamFct$new("grow_policy", default = "depthwise", levels = c("depthwise", "lossguide"), tags = "train"),
ParamInt$new("max_leaves", default = 0L, lower = 0L, tags = "train"),
ParamInt$new("max_bin", default = 256L, lower = 2L, tags = "train"),
ParamUty$new("callbacks", default = list(), tags = "train"),
ParamDbl$new("sketch_eps", default = 0.03, lower = 0, upper = 1, tags = "train"),
ParamDbl$new("scale_pos_weight", default = 1, tags = "train"),
ParamUty$new("updater", tags = "train"), # Default depends on the selected booster
ParamLgl$new("refresh_leaf", default = TRUE, tags = "train"),
ParamFct$new("feature_selector", default = "cyclic", levels = c("cyclic", "shuffle", "random", "greedy", "thrifty"), tags = "train"),
ParamInt$new("top_k", default = 0, lower = 0, tags = "train"),
ParamFct$new("predictor", default = "cpu_predictor", levels = c("cpu_predictor", "gpu_predictor"), tags = "train")
))
# param deps
ps$add_dep("tweedie_variance_power", "objective", CondEqual$new("reg:tweedie"))
ps$add_dep("print_every_n", "verbose", CondEqual$new(1L))
ps$add_dep("sample_type", "booster", CondEqual$new("dart"))
ps$add_dep("normalize_type", "booster", CondEqual$new("dart"))
ps$add_dep("rate_drop", "booster", CondEqual$new("dart"))
ps$add_dep("skip_drop", "booster", CondEqual$new("dart"))
ps$add_dep("one_drop", "booster", CondEqual$new("dart"))
ps$add_dep("tree_method", "booster", CondAnyOf$new(c("gbtree", "dart")))
ps$add_dep("grow_policy", "tree_method", CondEqual$new("hist"))
ps$add_dep("max_leaves", "grow_policy", CondEqual$new("lossguide"))
ps$add_dep("max_bin", "tree_method", CondEqual$new("hist"))
ps$add_dep("sketch_eps", "tree_method", CondEqual$new("approx"))
ps$add_dep("lambda", "booster", CondEqual$new("gblinear"))
ps$add_dep("lambda_bias", "booster", CondEqual$new("gblinear"))
ps$add_dep("alpha", "booster", CondEqual$new("gblinear"))
ps$add_dep("feature_selector", "booster", CondEqual$new("gblinear"))
ps$add_dep("top_k", "booster", CondEqual$new("gblinear"))
ps$add_dep("top_k", "feature_selector", CondAnyOf$new(c("greedy", "thrifty")))

# custom defaults
ps$values = list(nrounds = 1L, verbose = 0L)

super$initialize(
Expand Down
2 changes: 1 addition & 1 deletion man/LearnerClassifXgboost.Rd
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.