This repository contains the code (R) of the paper 'Debating Evil: Using Word Embeddings to Analyse Parliamentary Debates on War Criminals in the Netherlands' by Ralf Futselaar and Milan van Lange.
ABSTRACT: We are proposing a method to investigate changes in historical discourse by using large bodies of text and word embedding models (WEMs). As a case study, we investigate discussions in Dutch Parliament about the punishment of war criminals in the period 1945-1975. We will demonstrate how word embedding models, trained with Google’s Word2Vec algorithm, can be used to trace historical developments in parliamentary vocabulary through time.
The full-text online version of the paper can be found here.
To replicate our analysis and plot the results, you can download the trained .bin-files (vecs_1945-55.bin and vecs_1965-75.bin) from the 'releases' page and analyse them with the second script (2_analyse_plot_WEMs.R) in R or RStudio. Of course, you can also train new word embedding models based on your own dataset (a single, pre-processed plain .txt-file) with the first script (1_train_WEMs.R) and analyse them with the second (2_analyse_plot_WEMs.R). We recommend to use R version R 3.3.3 for Windows. In the scripts, we use the wordVectors package by Benjamin Schmidt and the magrittr forward-pipe operator package by Stefan Milton Bache and Hadley Wickham.
Our trained models are based on the PoliticalMashup and Dutch Parliamentary Proceedings dataset. We have used the complete data for the period 1945-1955 and 1965-1975. Because of file size limitations, we have only uploaded the trained WEMs based on this dataset. These .bin-files can de downloaded from this repository (see 'releases').
Source: Marx, Dr. M. (Informatics Institute, University of Amsterdam); Doornik, Dr. Ir. J. van (Informatics Institute, University of Amsterdam); Nusselder, BSc A. (Informatics Institute, University of Amsterdam); Buitinck, Msc L. (Informatics Institute, University of Amsterdam) (2012): Thematic collection: PoliticalMashup and Dutch Parliamentary Proceedings 1814-2013. DANS. https://doi.org/10.17026/dans-zg8-9x2v