EITM-based Item Identification

This section is a description of the Expert Informed Topic Model thematic classification of the plain text data.

The goal of this step is to identify the debates relevant to the master topic "the public role of religion in societal life".

First, an expert survey has been conducted that identified the relevant debates and the associated keywords that occured in each country.

After that, a topic model was trained and enriched with the expert information. The EITM-classifier produced a relevance score for each of the news items in the document corpus.

Finally, human coders thematically classified the scored document corpus in order to identify relevant documents and produce a sample for the content analysis step.