Machine learning algorithms in the use of mathematics didactic teaching and research

Goals: The research project is divided into two parts: One goal of the project is to write a deep learning algorithm that automatically evaluates freely formulated student answers to open questions from didactic lectures using supervised learning and classifies the respective answers as satisfactory (passed) or unsatisfactory (failed). In practical applications the evaluation of the text answers should be carried out in such a way that the algorithm recognizes possible learning difficulties and immediately afterwards provides automatic assistance in the sense of a digital feedback system and makes learning suggestions to the students. In addition to testing machine learning algorithms in the context of mathematics didactic teaching, a further objective is to investigate and describe the possibilities and limits within qualitative research processes using artificial intelligence. It is to be investigated whether unsupervised learning the "manual marking of text passages", as it is done for example in open coding within the framework of the "Grounded Theory" (Glaser & Strauss, 1967/2008, p. 63 f.), can be completely taken over by an ML-algorithm.

Use Methods: Artificial Neural Network, Data Mining, Machine Learning, Qualitative Research, Natural Language Processing, Text Mining, Quantitative Research Methods, Supervised Learning, Unsupervied Learning

Literatur: Glaser, B. G. & Strauss, A. L. (1967/2008): Grounded Theory. Strategien qualitativer Forschung, 2. Aufl., Mannheim: Huber.