Topic Modeling Workshop

On May 12th, 2020 from 9:00 a.m. to 12:30 p.m., a workshop on “Topic Modeling” will take place as part of “Mining and Modeling Text”. The workshop will take place online (via Zoom) and will be divided into two sessions, with a break in the middle (9: 00-10: 30 and 11: 00-12: 30). The reason for the workshop is that topic modeling will be part of the method repertoire in MiMoText; however, the workshop is open to anyone interested.

Topic Modeling is a quantitative method of text analysis with which one can discover recurring "topics" (in a very broad sense) in larger text collections. The topics are modeled as groups of words that appear again and again in similar contexts and between which there is a semantic relationship. In addition to "Word Embeddings", topic modeling is probably one of the most popular methods in the digital humanities, in which principles of distributional semantics come into play.

The first part of the workshop introduces both the basic idea of ​​topic modeling and at least some of the statistics behind the method. The second part is practice-oriented and explains how topic modeling can be implemented in Python. Basic knowledge of Python or another programming language is definitely helpful for the second part, but not a requirement.

Please register with Prof. Schöch by May 7th if you would like to take part in the workshop. Participation in the first part only is possible.

Some notes in advance:

(1) The presentation for the workshop is available online. The slides were created for a workshop in summer 2019 and are in English, but the workshop will be in German. (online, more recent) or (PDF, As of July 2019).

(2) Some preparation is necessary for the practical part. Notes on this can be found in the corresponding Github repository. There you will find installation instructions ("HOWTO" file) and a folder ("test") with instructions for testing the installation. Please install Python and all necessary libraries and test the installation before the workshop. If you have problems with the installation, please email us before the workshop!

Registration here