xiaodaigh / jlboost.jl Goto Github PK

View Code? Open in Web Editor NEW

68.0 4.0 6.0 215 KB

A 100%-Julia implementation of Gradient-Boosting Regression Tree algorithms

License: MIT License

Julia 100.00%

tree-boosting-algorithms tree catboost lightgbm gbdt gbrt machine-learning data-science xgboost

jlboost.jl's People

Contributors

Stargazers

Watchers

Forkers

statmixedml altre wallace-163 jordanmb margaeor jchartron

jlboost.jl's Issues

implement LightGBM style one var split algorithm

TagBot trigger issue

This issue is used to trigger TagBot; feel free to unsubscribe.

If you haven't already, you should update your TagBot.yml to include issue comment triggers.
Please see this post on Discourse for instructions and more details.

If you'd like for me to do this for you, comment TagBot fix on this issue.
I'll open a PR within a few hours, please be patient!

"Package JLBoost does not have Tables in its dependencies"

Same error in both master and release v0.1.5:

julia> using JLBoost
[ Info: Precompiling JLBoost [13d6d4a1-5e7f-472c-9ebc-8123a4fbb95f]
WARNING: both PrettyTables and MLJModelInterface export "matrix"; uses of it in module MLJBase must be qualified
ERROR: LoadError: LoadError: ArgumentError: Package JLBoost does not have Tables in its dependencies:
- If you have JLBoost checked out for development and have
  added Tables as a dependency but haven't updated your primary
  environment's manifest file, try `Pkg.resolve()`.
- Otherwise you may need to report an issue with JLBoost

DataFrames 0.20

(v1.3) pkg> add https://github.com/xiaodaigh/JLBoost.jl
  Updating git-repo `https://github.com/xiaodaigh/JLBoost.jl`
  Updating git-repo `https://github.com/xiaodaigh/JLBoost.jl`
 Resolving package versions...
ERROR: Unsatisfiable requirements detected for package DataFrames [a93c6f00]:
 DataFrames [a93c6f00] log:
 ├─possible versions are: [0.11.7, 0.12.0, 0.13.0-0.13.1, 0.14.0-0.14.1, 0.15.0-0.15.2, 0.16.0, 0.17.0-0.17.1, 0.18.0-0.18.4, 0.19.0-0.19.4, 0.20.0] or uninstalled
 ├─restricted to versions 0.19 by JLBoost [13d6d4a1], leaving only versions 0.19.0-0.19.4
 │ └─JLBoost [13d6d4a1] log:
 │   ├─possible versions are: 0.1.3 or uninstalled
 │   └─JLBoost [13d6d4a1] is fixed to version 0.1.3
 └─restricted to versions 0.20.0 by an explicit requirement — no versions left

parallel fit

Does it support Acceleration and Parallelism · MLJ ?

Do you have any idea on how to incorporate different weights for each datapoints, in the sense that datapoints with low weights should be less important for the model to fit than those with high weights?

For instance, by specifying the loss as WeightedLogitLogLoss(weights) etc.'

LossFunction.jl has some support for this, e.g.

value(LogitLogLoss(), Y, Y, AggMode.WeightedSum(ones(length(Y))))

monotonic constraint

Work with Tables.jl

Unable to install JLBoost.jl

Hey,
FYI there seem to be issues installing JLBoost now.

(v1.3) pkg> add JLBoost
  Updating registry at `C:\Users\azevelev\.julia\registries\General`
  Updating git-repo `https://github.com/JuliaRegistries/General.git`
 Resolving package versions...
ERROR: Unsatisfiable requirements detected for package JLBoost [13d6d4a1]:
 JLBoost [13d6d4a1] log:
 ├─possible versions are: 0.1.0-0.1.6 or uninstalled
 ├─restricted to versions * by an explicit requirement, leaving only versions 0.1.0-0.1.6
 ├─restricted by compatibility requirements with MLJBase [a7f614a8] to versions: [0.1.2-0.1.3, 0.1.6] or uninstalled, leaving only versions: [0.1.2-0.1.3, 0.1.6]
 │ └─MLJBase [a7f614a8] log:
 │   ├─possible versions are: [0.1.0-0.1.1, 0.2.0-0.2.6, 0.3.0, 0.4.0, 0.5.0, 0.6.0, 0.7.0-0.7.5, 0.8.0-0.8.4, 0.9.0-0.9.2, 0.10.0-0.10.1, 0.11.0-0.11.9] or uninstalled
 │   └─restricted to versions 0.11.9 by an explicit requirement, leaving only versions 0.11.9
 └─restricted by compatibility requirements with Tables [bd369af6] to versions: [0.1.0, 0.1.5] or uninstalled — no versions left
   └─Tables [bd369af6] log:
     ├─possible versions are: [0.1.0-0.1.15, 0.1.17-0.1.19, 0.2.0-0.2.11, 1.0.0-1.0.2] or uninstalled
     ├─restricted by compatibility requirements with MLJModels [d491faf4] to versions: [0.2.0-0.2.11, 1.0.0-1.0.2]
     │ └─MLJModels [d491faf4] log:
     │   ├─possible versions are: [0.1.0-0.1.1, 0.2.0-0.2.5, 0.3.0, 0.4.0, 0.5.0-0.5.9, 0.6.0-0.6.3, 0.7.0-0.7.2, 0.8.0-0.8.3] or uninstalled
     │   └─restricted to versions 0.8.3 by an explicit requirement, leaving only versions 0.8.3
     └─restricted by compatibility requirements with MLJScientificTypes [2e2323e0] to versions: 1.0.0-1.0.2
       └─MLJScientificTypes [2e2323e0] log:
         ├─possible versions are: [0.1.0-0.1.1, 0.2.0-0.2.1] or uninstalled
         └─restricted to versions 0.2.1 by an explicit requirement, leaving only versions 0.2.1

Also

(v1.3) pkg> add JLBoostMLJ
 Resolving package versions...
ERROR: Unsatisfiable requirements detected for package MLJ [add582a8]:
 MLJ [add582a8] log:
 ├─possible versions are: [0.1.0-0.1.1, 0.2.0-0.2.5, 0.3.0, 0.4.0, 0.5.0-0.5.9, 0.6.0-0.6.1, 0.7.0, 0.8.0, 0.9.0-0.9.3] or uninstalled
 ├─restricted to versions 0.9.3 by an explicit requirement, leaving only versions 0.9.3
 └─restricted by compatibility requirements with JLBoostMLJ [8b86df2c] to versions: 0.6.0-0.6.1 — no versions left
   └─JLBoostMLJ [8b86df2c] log:
     ├─possible versions are: 0.1.0 or uninstalled
     └─restricted to versions * by an explicit requirement, leaving only versions 0.1.0

I love your package & it would be awesome to use inside MLJ soon.
Check out EvoTrees.
They have a short program MLJ.jl that interfaces their package w/ MLJ.

(@v1.6) pkg> add DataFrames@v1.0.0
   Resolving package versions...
ERROR: Unsatisfiable requirements detected for package DataFrames [a93c6f00]:
 DataFrames [a93c6f00] log:
 ├─possible versions are: 0.11.7-1.0.0 or uninstalled
 ├─restricted to versions 1.0.0 by an explicit requirement, leaving only versions 1.0.0
 └─restricted by compatibility requirements with JLBoost [13d6d4a1] to versions: 0.19.0-0.22.7 — no versions left
   └─JLBoost [13d6d4a1] log:
     ├─possible versions are: 0.1.0-0.1.16 or uninstalled
     └─restricted to versions * by an explicit requirement, leaving only versions 0.1.0-0.1.16