WSDM Cup 2017: Triple scoring

Triple scoring (task 1) - WSDM Cup 2017

Introduction

We are a team of 4 from Complex Network Research Group (Murata Laboratory) - Tokyo Insitute of Technology. This repository is our submission to the 2017 WSDM Cup. In summary, the task is to predict human evaluation given name and country/profession.

Milestones

Preparation: 2016/09/26 - 2016/09/30

Objective

1. Setup personal computer to match each others. (Python 3.5.2, Tensorflow 0.10, scikit-learn 0.17.1, Anaconda virtual env, coding style, etc.)
2. Literature review. (Paper listing, reading, and discussion)
3. Competition score metric analysis.
4. Finalize and present possible approaches.

Daily log

1. 27th: List of natural language models; tripartite networks.
2. 28th: Paper reading; discussion about Skipgram and challenges in the training data and small set of labeled data.
3. 29th: Discussion about data preprocessing and prototype.
4. 30th: Meeting is skipped.