Studies in construction research are costly and time-consuming and therefore often only involve a small number of subjects. Therefore, although findings can often be shown analytically, they cannot be statistically validated. For methodological researchers, this represents a major challenge.
To solve this problem, the semi-automated evaluation of quantitative data collection methods, such as eye-tracking, offers great potential.

Task:

• Application of different machine learning approaches for classification to an existing eye-tracking dataset.
• Collection of further training data
• Runnable implementation of the developed algorithm, so that it can be applied in future studies as well

Profile:

• You study mechanical engineering or mechatronics
• You work purposefully, independently and on your own responsibility
• You have good programming skills in Python
• You have basic knowledge in Machine Learning

Then get in touch with me: christoph.zimmerer∂kit.edu