Name Nationality Classification with Recurrent Neural Networks

Implementation of Name Nationality Classification on Olympic Name Dataset

Paper will be published at IJCAI 2017 titled as: [Name Nationaltiy Classification with Recurrent Neural Networks]

• Jinhyuk Lee (jinhyuk_lee@korea.ac.kr)
• Hyunjae Kim (gamica@sogang.ac.kr)
• Miyoung Ko (gomi1503@korea.ac.kr)
• Donghee Choi (choidonghee@korea.ac.kr)
• Jaehoon Choi (jchoi@kono.ai)
• Jaewoo Kang (kangj@korea.ac.kr)

Requirements

• Python 3.4.3
• Tensorflow 1.0.1 (GPU enabled)
• Gensim

Directories

• data/ : Olympic Name Datasets including raw and cleaned version
• main.py : Running RNN model with adjustable hyperparameters
• model.py : Model structure for RNN-LSTM
• ops.py : Tensorflow ops used in model.py
• dataset.py : Dataset reading and experiment workflow
• char2vec.py : Making pretrained char2vec
• preprocess.py : Preprocessing crawled dataset
• utils.py : Utility for printing

Run the code

$ python main.py