xgboost在windows10系统下的安装

参考文档：

http://dnc1994.com/2016/03/installing-xgboost-on-windows/

https://stackoverflow.com/questions/43327020/xgboostlibrarynotfound-cannot-find-xgboost-library-in-the-candidate-path-did-y

一、安装步骤

1、编译

git clone --recursive https://github.com/dmlc/xgboost  
cd xgboost  
wget https://www.dropbox.com/s/y8myex4bnuzcp03/Makefile_win?dl=1
cp Makefile_win Makefile
cp make/mingw64.mk config.mk
mingw32-make

2、安装

cd python-package  
python setup.py install  

二、关键

1、Makefile_win采用第三方

    wget不好使的情况下， 可以这儿下载：

     https://www.dropbox.com/s/y8myex4bnuzcp03/Makefile_win?dl=0

2、采用gitbash，需要安装Git

3、--recursive参数

4、TDM-GCC下载安装

     http://tdm-gcc.tdragon.net/download

三、遇到的问题

1、问题一

     #python setup.py install

GBoostLibraryNotFound:Cannot find XGBoostLibraryin the candidate path, did you install compilers and run build.shin root path?
     解决：

     找到libxgboost.dll

     拷贝到xgboost/pythonp-package/xgboost中

     #mv libxgboost.dll  xgboost.dll

2、问题二

     #python setup.py install

    error: no lapack/blas resources found

   解决：

   1> 配置path路径：

         PATH=C:\Python27\Scripts

        升级pip命令

        # python -m pip install -U pip

   2> 安装scipy

        重启gitbash，不然依然无法识别pip命令

        #pip install scipy

四、intelij idea运行测试样例

1、样例

import xgboost as xgb

# read in data
dtrain = xgb.DMatrix('D:\\test\\xgboost\\demo\\data\\agaricus.txt.train')
dtest = xgb.DMatrix('D:\\test\\xgboost\\demo\\data\\agaricus.txt.test')
# specify parameters via map
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)

2、错误

      WindowsError: [Error 193] %1 is not a valid Win32 application

   解决：

     installing the python 64bit version. 

      查看python位数

      cmd窗口中输入python

        C:\Users\liuxq>python
Python 2.7.14 (v2.7.14:84471935ed, Sep 16 2017, 20:19:30) [MSC v.1500 32 bit (Intel)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>>

       下载64bit python重装：

        https://www.python.org/downloads/release/python-2714/

3、重新安装依赖软件

   numpy安装

   #pip install numpy

   

   scipy安装

   #pip install scipy




      xgboost安装： 

      liuxq@sh-liuxq MINGW64 /home/liuxq/xgboost/python-package (master)
      $ python setup.py install