I am working on machine learning algorithm development with imbalanced chemical structure-activity data. I need to install an interesting package, called “imbalanced-learn”. (see the documents: http://contrib.scikit-learn.org/imbalanced-learn/).
The Dependencies for imbalanced-learn is shown as follows:
imbalanced-learn is tested to work under Python 2.7 and Python 3.4, 3.5, and 3.6.
Additionally, to run the examples, you need matplotlib(>=2.0.0).
As such, I have to install latest version of scipy, numpy, scikit-learn, matplotlib.
- Download a latest version of Python 2.7 that binds with pip function
- Go to website: http://www.lfd.uci.edu/~gohlke/pythonlibs/
- numpy: numpy‑1.13.0+mkl‑cp27‑cp27m‑win32.whl
- scipy: scipy‑0.19.0‑cp27‑cp27m‑win32.whl
- Matplotlib: matplotlib‑2.0.2‑cp27‑cp27m‑win32.whl
You can install them via “pip install *.whl”
Alternatively, you can directly use “pip install numpy, scipy, Matplotlib”. In my case, it doesn’t work well with installing “scipy” on a 32-bit Windows machine.
3. Install scikit-learn
pip install -U scikit-learn
4. Install imbalanced-learn
pip install -U imbalanced-learn