Search
Program Calendar
Browse By Day
Browse By Time
Browse By Person
Browse By Room
Browse By Division
Browse By Session Type
Search Tips
Personal Schedule
Sign In
Latent Dirichlet allocation (LDA) topic models are increasingly being used in communication research. Yet, the reliability and validity of the approach have received little attention thus far. In applying LDA to textual data, researchers need to tackle at least four major challenges affecting these criteria: (a) the appropriate preprocessing of the text collection; (b) the adequate selection of model parameters including the number of topics to be generated; (c) the evaluation of the model’s reliability, and (d) the proceeding of validly interpreting the resulting topics. We review the research literature dealing with these questions and report own experiences before we propose a methodology which approaches these challenges. Our overall goal is to make LDA topic modeling more accessible to communication researchers and to ensure compliance with disciplinary standards. We demonstrate the value of our approach with sample data from an ongoing research project.
Daniel Maier, Free U of Berlin
Annie Waldherr
Peter Miltner, FU Berlin
Gregor Wiedemann, U of Leipzig
Andreas Niekler, U of Leipzig
Gerhard Heyer, U of Leipzig
Alexa Keinert, Freie U Berlin
Barbara Pfetsch, Freie Universitaet Berlin
Thomas Haeussler, University of Bern
Ueli Reber, U Bern
Hannah Schmid-Petri, U of Passau
Silke Adam, Universität Bern, IKMB