Replication Code for 'Moving On' - Investigating Inventors' Ethnic Origins Using Supervised Learning

This repository contains the code to replicate the paper Moving On - Investigating Inventors' Ethnic Origins Using Supervised Learning.

Repository Structure

Datasets that were created in this analysis can be found in the folder 00_data_and_model. The trained and tuned LSTM classification model used for the analysis in this paper is stored in this folder as well and can be accessed under 00_data_and_model/model/name_origin_lstm.h5. The folder 01_create_training_dataset contains replication files used to construct the dataset of labeld names used to train the LSTM classification model. 02_model_training features the code to train the LSTM classifier. Lastly, the code for the descriptive analysis (using a random subsample of the paper'sb dataset) can be found in the folder 03_inventor_composition_analysis

Dependencies

Python (3.7)

• joblib==1.0.1
• matplotlib==3.3.1
• numpy==1.19.2
• pandas==1.1.3
• pyreadr==0.3.5
• scikit-learn==0.23.2
• scipy==1.4.1
• tensorflow==2.2.0
• xgboost==0.90

Installing a virtual environment using the environment.yml or requirements.txt files is recommended.

R (4.0.1)

• tidyverse
• data.table
• reticulate
• tensorflow
• keras
• stringi
• jsonlite
• countrycode
• viridis

References & Contact

Niggli, M. (2022), 'Moving On' -- Investigating Inventors' Ethnic Origins Using Supervised Learning, arXiv:2201.00578

If you have questions, please contact [email protected].