Competence Identification and Classification of Computer Science Teaching Materials

05. Juli 2017, 10:00 , 11:30

Veranstalter:  Dipl.-Inform. Jörn Syrbe, Universität Oldenburg
Ort:  OFFIS, Escherweg 2, Raum F02

Abstract:
The standards for computer science education of the “German Informatics Society” (Gesellschaft für In-
formatik e.V. – GI) and the “Curriculum for Lower Secondary Education in Informatics in Lower Saxony”
indicate which competences should be taught. In order to teach these competences, teachers need to create
materials by themselves or use existing materials like commercial textbooks or teaching materials from the
Internet (or from the intranet of schools). Teachers that are searching through these online materials have to
evaluate the relevance of each one. Because of the varying lesson topics, lesson ideas and quality of the
materials, the whole process require a considerable amount of time and effort. In addition to this, it is chal-
lenging for teachers to determine which of these materials encourage, improve and foster which kind compe-
tences.
The current approach strives to identify competences within materials by classifying them by a set of data
retrieval, text mining and artificial intelligence techniques. These techniques are stemming algorithms based
on n-grams, which are used to identify the co-occurence of the adjacency of words within sentences and texts.
The classification is based on the definitions of computer science competences for lower secondary education
from the GI and the Lower Saxony curriculum. The material corpus contains German text-based materials
from the internet and publishing companies (e.g. Cornelsen Verlag, Ernst Klett Verlag and Herdt-Verlag).
The classification handles the numerous German language levels within the materials and standards by using
similarity measures, differential analysis and various German dictionaries and thesauruses. The outcome of
this approach is a classification of German computer science education materials by competences.
Betreuer: Prof. Dr. Ira Diethelm