Window下XGBoost安装

本文主要是对XGBoost安装进行介绍。

在正文开始前，需要安装一些前置软件如下：

 - Git
 - MINGW

下面开始正文介绍。

1.首先创建XGBoost安装目录文件，然后在该目录下面启动git bash.

E:\Xgboost

2.然后执行如下命令。

$ git clone --recursive https://github.com/dmlc/xgboost
$ cd xgboost
$ git submodule init
$ git submodule update

3.下载MinGW-W64，然后安装。

4.这里需要修改Architecture值为x86_64，因为我的本机是64位操作系统。

5.这里安装在我们前面创建的文件目录下面。

6.然后将mingw64\bin配置到path中，然后重启bash，检验是否安装成功.

$ which mingw32-make
/e/Xgboost/mingw-w64/x86_64-7.1.0-posix-seh-rt_v5-rev1/mingw64/bin/mingw32-make

7.取别名

$ alias make='mingw32-make'

8.逐个编译各个模块

$ cd xgboost/

$ pwd
/e/Xgboost/xgboost

$ cd dmlc-core
$ make -j4
$ cd ../rabit
$ make lib/librabit_empty.a -j4
$ cd ..
$ cp make/mingw64.mk config.mk
$ make -j4

9.然后打开xgboost的python-package

$ cd python-package/

$ pwd
/e/Xgboost/xgboost/python-package

10.安装

/e/Xgboost/xgboost/python-package/python setup.py install

11.打开cmd，启动python,然后在命令行中加入如下内容。

import os

mingw_path = 'E:\\Xgboost\\mingw-w64\\x86_64-7.1.0-posix-seh-rt_v5-rev1\\mingw64\\bin'

os.environ['PATH'] = mingw_path + ';' + os.environ['PATH']

12.验证是否安装成功.

>>> import xgboost as xgb

>>> dtrain = xgb.DMatrix('E:/data/agaricus.txt.train')
[19:30:53] 490x127 matrix with 10780 entries loaded from E:/data/agaricus.txt.train

>>> dtest = xgb.DMatrix('E:/data/agaricus.txt.test')
[19:31:19] 177x127 matrix with 3894 entries loaded from E:/data/agaricus.txt.test

>>> param = {'max_depth':2, 'eta':1, 'silent':1, 'objective':'binary:logistic' }
>>> num_round = 2
>>> bst = xgb.train(param, dtrain, num_round)
>>> # make prediction
... preds = bst.predict(dtest)
>>>
>>> preds
array([ 0.0432693,  0.9445464,  0.0432693,  0.0432693,  0.0432693,
        0.0432693,  0.9445464,  0.0432693,  0.9445464,  0.0432693,
        0.9445464,  0.0432693,  0.0432693,  0.0432693,  0.0432693,
        0.0432693,  0.0432693,  0.9445464,  0.0432693,  0.0432693,
        0.0432693,  0.0432693,  0.0432693,  0.0432693,  0.0432693,
        0.9445464,  0.0432693,  0.0432693,  0.0432693,  0.0432693,
                    .........

参考文献：

https://www.ibm.com/developerworks/community/blogs/jfp/entry/Installing_XGBoost_For_Anaconda_on_Windows?lang=zh
https://xgboost.readthedocs.io/en/latest/build.html
https://xgboost.readthedocs.io/en/latest/get_started/
https://github.com/dmlc/xgboost/tree/master/demo/data