Quantitative analysis of large amounts of journalistic texts using topic modelling

摘要: The huge collections of news content which have become available through digital technologies both enable and warrant scientific inquiry, challenging journalism scholars to analyse unprecedented amounts texts. We propose Latent Dirichlet Allocation (LDA) topic modelling as a tool face this challenge. LDA is cutting edge technique for analysis, designed automatically organize large archives documents based on latent topics, measured patterns word (co-)occurrence. explain how works, different choices by the researcher affect results can be meaningfully interpreted. To demonstrate its usefulness research, we conducted case study New York Times coverage nuclear technology from 1945 present, partially replicating Gamson Modigliani. This shows that useful analysing trends in relatively quickly.