For training boosted tree models, there are 2 parameters used for choosing algorithms,
tree_method. XGBoost has 4 builtin tree methods, namely
gpu_hist. Along with these tree methods, there
are also some free standing updaters including
sync. The parameter
updater is more primitive than
as the latter is just a pre-configuration of the former. The difference is mostly due to
historical reasons that each updater requires some specific configurations and might has
missing features. As we are moving forward, the gap between them is becoming more and
more irrevelant. We will collectively document them under tree methods.
Exact means XGBoost considers all candidates from data for tree splitting, but underlying the objective is still interpreted as a Taylor expansion.
exact: Vanilla tree boosting tree algorithm described in reference paper. During each split finding procedure, it iterates
over every entry of input data. It’s more accurate (among other greedy methods) but
slow in computation performance. Also it doesn’t support distributed training as
XGBoost employs row spliting data distribution while
exact tree method works on a
sorted column format. This tree method can be used with parameter
exact tree method is slow in performance and not scalable, we often employ
approximated training algorithms. These algorithms build a gradient histogram for each
node and iterate through the histogram instead of real dataset. Here we introduce the
implementations in XGBoost below.
grow_local_histmaker updater: An approximation tree method described in reference
paper. This updater is rarely used in practice so
it’s still an updater rather than tree method. During split finding, it first runs a
weighted GK sketching for data points belong to current node to find split candidates,
using hessian as weights. The histogram is built upon this per-node sketch. It’s
exact in some applications, but still slow in computation.
approx tree method: An approximation tree method described in reference paper. Different from
grow_local_histmaker, it runs
sketching before building each tree using all the rows (rows belonging to the root)
instead of per-node dataset. Similar to
grow_local_histmaker updater, hessian is
used as weights during sketch. The algorithm can be accessed by setting
hist tree method: An approximation tree method used in LightGBM with slight
differences in implementation. It runs sketching before training using only user
provided weights instead of hessian. The subsequent per-node histogram is built upon
this global sketch. This is the fastest algorithm as it runs sketching only once. The
algorithm can be accessed by setting
gpu_hist tree method: The
gpu_hist tree method is a GPU implementation of
hist, with additional support for gradient based sampling. The algorithm can be
accessed by setting
Some objectives like
reg:squarederror have constant hessian. In this case,
gpu_hist should be preferred as weighted sketching doesn’t make sense with constant
weights. When using non-constant hessian objectives, sometimes
approx yields better
accuracy, but with slower computation performance. Most of the time using
max_bin can achieve similar or even superior accuracy while maintaining
good performance. However, as xgboost is largely driven by community effort, the actual
implementations have some differences than pure math description. Result might have
slight differences than expectation, which we are currently trying to overcome.
Pruner: It prunes the built tree by
pruner is usually
used as part of other tree methods.
Refresh: Refresh the statistic of bulilt trees on a new training dataset.
Sync: Synchronize the tree among workers when running distributed training.
2 Updaters were removed during development due to maintainability. We describe them here
solely for the interest of documentation. First one is distributed colmaker, which was a
distributed version of exact tree method. It required specialization for column based
spliting strategy and a different prediction procedure. As the exact tree method is slow
by itself and scaling is even less efficient, we removed it entirely. Second one is
skmaker. Per-node weighted sketching employed by
grow_local_histmaker is slow,
skmaker was unmaintained and seems to be a workaround trying to eliminate the
histogram creation step and uses sketching values directly during split evaluation. It
was never tested and contained some unknown bugs, we decided to remove it and focus our
resources on more promising algorithms instead. For accuracy, most of the time
gpu_hist are enough with some parameters tunning, so removing
them don’t have any real practical impact.