Automated Topic Model Extraction using Latent Dirichlet Allocation on the Example of Reviews on Car Manufacturers

Thesis Type Master
Thesis Status
Student Franziska Scharpf
Thesis Supervisor

The aim of this master thesis is to filter the most relevant topics and terms out of a set
of unstructured data, which refer to customer feedback of car manufacturers, and
to present them graphically. For this purpose, the following research question is posed:
How can the most relevant terms and topics be filtered out of documents with the help
of Latent Dirichlet Allocation?