Collection of machine learning utilities for PicNet and Predict Bench

PicNet and Predict Bench provide predictive analytics services and products like Centazio. These products and services are supported by this library that combines best in breed libraries, implementations, algorithms and utilities that help us provice machine learning services at speed.

See http://www.picnet.com.au for more details

Instructions:

• Python 2:
  • Check out a submodule to this lib name it ml
  • Create a <project_name>_utils.py file with project wide utilties
  • In <project_name>_utils.py add "from ml import *"
• Python 3:
  • Expectes a folder structure as follows:
    • src
      • utils.py (with from ml import *)
      • script01.py (with import src.utils)
    • ml [git submodule to this lib]
  • To run a script use python -m src.script01
  • Or in ipython import src.utils to get going
• Jupyter Notebook
  • ml will need to live in the src directory
  • "from ml import *"

This will inject all the required libraries into your environment including:

• pandas (as pd)
• numpy (as np)
• scipy
• sklearn
• all utiltiy functions in misc.py
• all pandas extensions defined in pandas_extensions

License: MIT Author: Guido Tapia - [email protected]