OSOM A method for building overlapping topological maps
OSOM: A method for building overlapping topological maps PRESENTER : CHUANG, KAI-TING AUTHORS : GUILLAUME CLEUZIOU* 2013, PRL Intelligent Database Systems Lab
Outlines n Motivation n Objectives n Methodology n Experiments n Conclusions n Comments Intelligent Database Systems Lab
Motivation • Overlapping clustering solutions extract data organizations that are more fitted to the input data than crisp clustering solutions. • Unsupervised neural networks bring efficient solutions to visualize class structures. Intelligent Database Systems Lab
Objectives • We present the algorithm O-SOM that uses both an overlapping variant of the k-means clustering algorithm and the well known Kohonen approach, in order to build overlapping topologic maps. • To solve problems that are recurrent in overlapping clustering: number of clusters, complexity of the algorithm and coherence of the overlaps. Intelligent Database Systems Lab
Methodology-Framework Intelligent Database Systems Lab
Methodology OSOM Intelligent Database Systems Lab
Methodolog-fast-osom Intelligent Database Systems Lab
Experiment-dataset Intelligent Database Systems Lab
Experiment-evaluation framework Intelligent Database Systems Lab
Experiment results Intelligent Database Systems Lab
Experiment-Topological evaluation Intelligent Database Systems Lab
Conclusions • Ensure the algorithm to converge and then bring solutions to the motivations mentioned: limited complexity, topological correctness, etc. Intelligent Database Systems Lab
Comments • Advantages – The OSOM is simple method. • Applications – Topological maps. Intelligent Database Systems Lab
- Slides: 13