Thesis supervisor: József Dombi
Location of studies (in Hungarian): SZTE Abbreviation of location of studies: SZTE
Description of the research topic:
Clustering is a part of the supervisor –free machine learning method. During the procedure the original database is divided into smaller clusters.
It is well known if our aim is to carry out a supervised learning procedure, the running time of this procedure usually increases exponentially with the size of the database. Hence clustering can be a useful tool here. We can summarized the principle like so: divide and conquer. Fortunately all the elements of one cluster belongs to one class, i.e. the elements of these clusters have already been learned.
Task to be valued:
1. Developing the kind of clustering algorithms that handles the machine learning tasks.
2. Developing clustering algorithms where the member of the clusters is automatically determined
3. Studying fuzzy c-means algorithms and developing new variants
4. Coupling fuzzy regression tree to a fuzzy clustering algorithm
5. Coupling a time series analysis to a clustering algorithm.
Bibliography: Babuska: Fuzzy modeling for control; Baldwin: Time series modeling; Fridman: Multivariate regression splines