Unsupervised Learning is a part of the field of machine learning. Main goals are to detect structure in unknown data and classify data according to this structure. The methods can be applied to data from almost any field, but we'll concentrate on biological or biomedical data.
Some important algorithms are the k-nearest neighbour classificator, k-means clustering, hierarchical agglomerative clustering, Self-Organizing Maps, Neural Gas, and Spectral Clustering. To assess the quality of clusters or classification results, different evaluation methods are used which will also be a topic of this seminar.
In addition, different distance measures and dimension reduction methods like Principle Component Analysis, Single Value Decomposition, and Sammon Mapping are to be presented.
Rhythmus | Tag | Uhrzeit | Format / Ort | Zeitraum |
---|
Studiengang/-angebot | Gültigkeit | Variante | Untergliederung | Status | Sem. | LP | |
---|---|---|---|---|---|---|---|
Bioinformatik und Genomforschung / Bachelor | (Einschreibung bis SoSe 2011) | Modul 10; Musterkennung; Mathematische Methoden | Wahlpflicht | 6. | 3 | unbenotet | |
Kognitive Informatik / Bachelor | (Einschreibung bis SoSe 2011) | Neuronale Netze und Lernen | Pflicht | 6. | 3 | unbenotet | |
Naturwissenschaftliche Informatik / Bachelor | (Einschreibung bis SoSe 2011) | Vert Informatik | Wahlpflicht | 6. | 3 | unbenotet | |
Naturwissenschaftliche Informatik / Diplom | (Einschreibung bis SoSe 2004) | Biologie; Physik; Robotik | HS | ||||
Naturwissenschaftliche Informatik / Diplom | (Einschreibung bis SoSe 2004) | ME; NNet; BioI | HS |