EvoTrees

Documentation	CI Status	DOI

A Julia implementation of boosted trees with CPU and GPU support. Efficient histogram based algorithms with support for multiple loss functions (notably multi-target objectives such as max likelihood methods).

R binding available.

Installation

Latest:

julia> Pkg.add(url="https://github.com/Evovest/EvoTrees.jl")

From General Registry:

julia> Pkg.add("EvoTrees")

Performance

Data consists of randomly generated Matrix{Float64}. Training is performed on 200 iterations.
Code to reproduce is available in benchmarks/regressor.jl.

• Run Environment:
  • CPU: 12 threads on AMD Ryzen 5900X
  • GPU: NVIDIA RTX A4000
  • Julia: v1.10.8
• Algorithms
  • XGBoost: v2.5.1 (Using the hist algorithm)
  • EvoTrees: v0.17.0

CPU:

nobs	nfeats	max_depth	train_evo	train_xgb	infer_evo	infer_xgb
100k	10	6	0.4	0.7	0.0	0.0
100k	10	11	5.8	1.1	0.1	0.1
100k	100	6	1.2	1.4	0.1	0.1
100k	100	11	18.3	3.5	0.1	0.2
1M	10	6	2.5	6.3	0.3	0.3
1M	10	11	11.6	8.0	0.7	0.6
1M	100	6	6.5	14.7	0.7	1.3
1M	100	11	33.4	19.0	1.2	1.7
10M	10	6	28.6	86.7	3.9	2.9
10M	10	11	66.6	113.0	6.9	6.3
10M	100	6	74.2	151.0	6.6	14.2
10M	100	11	198.0	192.0	12.2	17.8

GPU:

nobs	nfeats	max_depth	train_evo	train_xgb	infer_evo	infer_xgb
100k	10	6	1.14	0.28	0.01	0.02
100k	10	11	17.56	1.29	0.01	0.02
100k	100	6	1.75	0.61	0.04	0.14
100k	100	11	32.62	3.21	0.04	0.17
1M	10	6	2.27	0.96	0.05	0.15
1M	10	11	27.10	2.73	0.06	0.19
1M	100	6	3.71	2.89	0.35	1.37
1M	100	11	45.50	7.90	0.37	1.63
10M	10	6	9.11	7.46	0.53	1.73
10M	10	11	46.86	13.13	0.59	1.76
10M	100	6	22.74	28.32	3.43	14.77
10M	100	11	80.63	52.68	3.50	17.88

MLJ Integration

See official project page for more info.

Quick start with internal API

A model configuration must first be defined, using one of the model constructor:

• EvoTreeRegressor
• EvoTreeClassifier
• EvoTreeCount
• EvoTreeMLE

Model training is performed using fit_evotree. It supports additional keyword arguments to track evaluation metric and perform early stopping. Look at the docs for more details on available hyper-parameters for each of the above constructors and other options training options.

Matrix features input

using EvoTrees
using EvoTrees: fit

config = EvoTreeRegressor(
    loss=:mse, 
    nrounds=100, 
    max_depth=6,
    nbins=32,
    eta=0.1)

x_train, y_train = rand(1_000, 10), rand(1_000)
m = fit(config; x_train, y_train)
preds = m(x_train)

DataFrames input

When using a DataFrames as input, features with elements types Real (incl. Bool) and Categorical are automatically recognized as input features. Alternatively, fnames kwarg can be used to specify the variables to be used as features.

Categorical features are treated accordingly by the algorithm: ordered variables are treated as numerical features, using ≤ split rule, while unordered variables are using ==. Support is currently limited to a maximum of 255 levels. Bool variables are treated as unordered, 2-levels categorical variables.

dtrain = DataFrame(x_train, :auto)
dtrain.y .= y_train
m = fit(config, dtrain; target_name="y");
m = fit(config, dtrain; target_name="y", fnames=["x1", "x3"]);

Feature importance

Returns the normalized gain by feature.

features_gain = EvoTrees.importance(m)

Plot

Plot a given tree of the model:

plot(m, 2)

Note that 1st tree is used to set the bias so the first real tree is #2.

Save/Load

EvoTrees.save(m, "data/model.bson")
m = EvoTrees.load("data/model.bson");