Text Mining von Vokabeln für architektonische und akustische Bewertungen
A very long time ago, language prevailed to communicate perception and feelings. For architects, who aim to meet the clients’ expectations, it is thus especially important to understand how people verbalize their perception of a designed space. In our research on person-focused analysis of architectural design, we more specifically explore visual and auditory experience in these environments. Therefore, we are aiming to understand how we can relate quantifiable measures to individual perception.
As a first step, it is thus essential to collect the words that can be found being used to describe architecture and its acoustics. This vocabulary could be found in scientific as well as in popular literature. The idea is to structure them and cluster related attributes. We are interested in understanding which words are most popular, what synonyms are used, as well as how specific words are understood.
The approach we propose in this thesis is to find a suitable text mining method that can then be applied to find the respective vocabulary. They can afterwards be transferred to a structured database of attribute descriptors. Finally, they can be compared with descriptors gathered from subjective judgements during audio-visual experiments.
- Basic knowledge & interest in programming
- Fundamentals of machine learning
- Independent work attitude
- Enthusiasm for language & texts