python開發學習記錄--Numpy、Scipy、Matplotlib、Scikit-learn等庫的安裝
其實很簡單，在git bash中，輸入：
[email protected] MINGW64 /d/SoftWare/Python/Python37/Scripts
$ pip install numpy
$ pip install matplotlib
看到Successfully installed，說明OK。
注：前面的方法，在安裝時，遇到了問題，所以請參考下面的方法：
實際在使用Scikit-learn時，出錯了，後來發現是numpy+mkl和Scipy沒安裝。
安裝Scikit-learn之前，先裝Numpy\Scipy\Matplotlib包，再安裝Scikit-Learn。
去到官網下載相應版本：
http://www.lfd.uci.edu/~gohlke/pythonlibs
其中cp27表示CPython 2.7版本，cp34表示CPython 3.4，win_arm64指的是64位版本。
參考：https://blog.csdn.net/hujiuding/article/details/78309309
為了避免衝突， numpy scipy matplotlib 等都應該在官網下載。
值得注意的是：
Scikit-learn integrates classic machine learning algorithms.
Requires
Scikit-Learn下載的小節，提示說需要 numpy+mkl.
如果你安裝的numpy不含mkl，那麼Scikit-Learn可能會報錯。
因此，建議是在官網都下載相應檔案。
另外，numpy+mkl 比較大，我的瀏覽器下載速度很慢，我是用百度雲盤先離線下載後，再下到本地的。
雲工具可參考：https://blog.csdn.net/lemonix7/article/details/
另外值得注意的是，在安裝numpy時，
下載到的檔名是：
numpy-1.15.1 mkl-cp37-cp37m-win_amd64.whl
而官方的檔名是（多了一個加號）
numpy‑1.15.1+mkl‑cp37‑cp37m‑win_amd64.whl
自己記得改一下名。
安裝成功後，以下import就不會報錯了。
from sklearn import linear_model最終的安裝日誌如下，供參考：
[email protected] MINGW64 /d/SoftWare/Python/Python37/Scripts/lib
$ ll
total 250904
-rw-r--r-- 1 ALI 197121 8734219 九月 8 23:55 matplotlib-2.2.3-cp37-cp37m-win_amd64.whl
-rw-r--r-- 1 ALI 197121 230383926 九月 9 01:18 numpy-1.15.1+mkl-cp37-cp37m-win_amd64.whl
-rw-r--r-- 1 ALI 197121 4481636 九月 8 13:55 scikit_learn-0.19.2-cp37-cp37m-win_amd64.whl
-rw-r--r-- 1 ALI 197121 13313205 九月 8 23:54 scipy-1.1.0-cp37-cp37m-win_amd64.whl
-rw-r--r-- 1 ALI 197121 0 九月 8 23:16 注：該資料夾是用於存放從官網下載的whl檔案.txt
[email protected] MINGW64 /d/SoftWare/Python/Python37/Scripts/lib
$ pip install numpy-1.15.1+mkl-cp37-cp37m-win_amd64.whl
Processing d:\software\python\python37\scripts\lib\numpy-1.15.1+mkl-cp37-cp37m-win_amd64.whl
Installing collected packages: numpy
Successfully installed numpy-1.15.1+mkl
[email protected] MINGW64 /d/SoftWare/Python/Python37/Scripts/lib
$ pip install matplotlib-2.2.3-cp37-cp37m-win_amd64.whl
Processing d:\software\python\python37\scripts\lib\matplotlib-2.2.3-cp37-cp37m-win_amd64.whl
Requirement already satisfied: kiwisolver>=1.0.1 in d:\software\python\python37\lib\site-packages (from matplotlib==2.2.3) (1.0.1)
Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in d:\software\python\python37\lib\site-packages (from matplotlib==2.2.3) (2.2.0)
Requirement already satisfied: python-dateutil>=2.1 in d:\software\python\python37\lib\site-packages (from matplotlib==2.2.3) (2.7.3)
Requirement already satisfied: cycler>=0.10 in d:\software\python\python37\lib\site-packages (from matplotlib==2.2.3) (0.10.0)
Requirement already satisfied: numpy>=1.7.1 in d:\software\python\python37\lib\site-packages (from matplotlib==2.2.3) (1.15.1+mkl)
Requirement already satisfied: pytz in d:\software\python\python37\lib\site-packages (from matplotlib==2.2.3) (2018.5)
Requirement already satisfied: six>=1.10 in d:\software\python\python37\lib\site-packages (from matplotlib==2.2.3) (1.11.0)
Requirement already satisfied: setuptools in d:\software\python\python37\lib\site-packages (from kiwisolver>=1.0.1->matplotlib==2.2.3) (40.2.0)
Installing collected packages: matplotlib
Successfully installed matplotlib-2.2.3
[email protected] MINGW64 /d/SoftWare/Python/Python37/Scripts/lib
$ pip install scipy-1.1.0-cp37-cp37m-win_amd64.whl
Processing d:\software\python\python37\scripts\lib\scipy-1.1.0-cp37-cp37m-win_amd64.whl
Requirement already satisfied: numpy>=1.8.2 in d:\software\python\python37\lib\site-packages (from scipy==1.1.0) (1.15.1+mkl)
Installing collected packages: scipy
Successfully installed scipy-1.1.0
[email protected] MINGW64 /d/SoftWare/Python/Python37/Scripts/lib
$ pip install scikit_learn-0.19.2-cp37-cp37m-win_amd64.whl
Processing d:\software\python\python37\scripts\lib\scikit_learn-0.19.2-cp37-cp37m-win_amd64.whl
Installing collected packages: scikit-learn
Successfully installed scikit-learn-0.19.2
[email protected] MINGW64 /d/SoftWare/Python/Python37/Scripts/lib
$