Organisation of knowledge in large datasets: clustering algorithms and strategies
Visit at the Faculty of Forestry, University of Agriculture, Cracow, Poland
July 3-13, 2016
Host: Prof. P. Wezyk
Data mining is an ubiquitous theme that is present in many fields of science: surely in pattern recognition, classification of images, detecting communities in networks, and even decision trees and neural networks. Dr. Wezyk at the University of Agriculture, Cracow, Poland is actually engaged in a challende launched by the European Commission in the mapping of high resolution satellite data and Lidar data. For “Big Data” a conceptual approach must be proven to have a practical impact.
The detailed STSM report about this topic can be found here: PDF.
Summing up, the visit has shown to be effective for a better integration of techniques and methods that have been developed for examining data in quite different disciplines. While images from satellite have always constituted a large source of data, the new social networks platforms, communication platforms, financial databases, etc. allow to consider the extraction of information
from large datasets also in socio-economic environments. On the reversal way, not-strict classification used in complex networks clustering add new classification techniques to the already existing fuzzy set-based classification already known in data mining.
During the visit, there has been some interaction with students and postdocs. On July 11th, a crash course on MATLAB on main topics and key Toolboxes (interpolation, optimization, visualization), necessary for working on the subject. The pictures in the Appendix 2 show the seminar room. In conclusion, the scientific interchange among different fields has been quite fruitful, and it has shown potentialities ok forward for future work on the topic. In conclusion, this first visit, and the exchange of information among different fields has shown potentialities that may be expanded in the future.
Cracow, July 13th, 2016
Giulia Rotundo
