Schema Inference and Machine Learning (SIML)
Internally funded project
Acronym: SIML
Start date : 01.08.2018
Project details
Short description
Within the project SIML (Schema Inference and Machine Learning), methods of topological data analysis and unsupervised learning are combined, applied and further developed in order to derive a conceptual schema from unstructured, multivariant data.
Scientific Abstract
Within the framework of the project SIML (Schema Inference and Machine Learning), unstructured and semi-structured data are to be used to generate information from which a partial conceptual schema can be derived. Methods of topological data analysis (TDA) are used in combination with machine learning techniques to automate this as far as possible. In particular, we are interested in a stable, persistent form of natural data when using unsupervised learning methods. As a core concept, functional dependencies after data processing are to be investigated, with the help of which a suitable schema can then be defined. There are parallels and differences for time series and persistent data, which are also to be worked out.
The motivation of the work is to prove that schemata have a natural geometric structure in the form of a simplicial complex which can be investigated or made visible by topological methods.
