Master's Thesis David Koller

Interactive Analysis of a Corpus of General Terms and Conditions for Variability Modeling

Abstract

The Digitalization of knowledge and information transforms most industries, including the legal domain. With retail E-commerce sales continuously growing worldwide, many new digital marketplaces and online shops are created, all of which need to specify their own general terms and conditions to conduct business online. Unfortunately, there is no legally binding standard for general terms and conditions, which leads to high variety between retailers with regards to content, form and lawfulness of their terms and conditions.

This research work will analyze a corpus of general terms and conditions (“Allgemeine Geschäftsbedingungen”) form German online shops. First, clauses from terms and conditions will be clustered based on their semantic content by implementing an interactive Machine Learning (iML) algorithm. Several approaches are possible, starting purely unsupervised before interaction with a human expert or providing desirable classes right from the start. Afterwards by using variability modeling the resulting clusters will be analyzed to identify properties of the clauses, like mandatory and optional clauses, dependencies or mutual exclusivity, and lawfulness of clauses.