Introduction

The repo contains different implementations of the Kaggle Whats-cooking challenge (https://www.kaggle.com/c/whats-cooking)

Installation and Configuration

Install Anaconda Open Data Science Platform (Linux)

1. Download and Install Anaconda

   Download Anaconda distribution from https://www.continuum.io/downloads

   Run Anaconda3-4.0.0-Linux-x86_64.sh to install (without sudo)

2. Create a virtual environment

   Run conda create --name <env_name> --file </path/to/file/>requirements.txt Type “y” for “yes.”

3. Change environments (activate)

   source activate <env_name>

For more information about Anaconda, visit https://www.continuum.io/

conda documentation can be found at http://conda.pydata.org/docs/index.html

Applications

• main_nn.py and funcsEx04.py - A neural network based on Machine Learning course on Coursera https://www.coursera.org/learn/machine-learning. Rewritten in Python by royshoo (https://github.com/royshoo/mlsn) with my modification for the Whats-cooking task.
• main_sklearn_svm.py - OneVsRestClassifier on SVM core with the scikit-learn framework (http://scikit-learn.org/).
• main_sklearn_lr.py - OneVsRestClassifier on Linear Regression core.
• main_tf_gd.py - !Experimental! Tensorflow framework gradient descent (https://www.tensorflow.org/).
• utils.py - utils methods for work with the Whats-cooking dataset.

All rights belong to their owners.