Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

SOE: Fachverband Physik sozio-ökonomischer Systeme

SOE 5: Focus Session: Big Data

SOE 5.3: Hauptvortrag

Montag, 26. März 2012, 16:00–16:30, HE 101

Embedding high dimensional data on networks — •Tiziana DiMatteo — King's College London

In this talk I will introduce a graph-theoretic approach to extract clusters and hierarchies in complex data- sets in an unsupervised and deterministic manner, without the use of any prior information [1,2]. This is achieved by building topologically embedded networks containing the subset of most significant links and analyzing the network structure. For a planar embedding [3] this method provides both the intra-cluster hierarchy, which describes the way clusters are composed, and the inter-cluster hierarchy which describes how clusters gather together. I will discuss performance, robustness and reliability of this method by investigating several synthetic data-sets finding that it can outperform significantly other established approaches. Applications to financial data-sets show that industrial sectors and specific activities can be extracted and meaningfully identified from the analysis of the collective fluctuations of prices in an equity market.

[1] Won-Min Song, T. Di Matteo, T. Aste, Discrete Applied Mathematics 159 (2011) 2135. [2] Won-Min Song, T. Di Matteo, T. Aste, "Hierarchical information clustering by means of topologically embedded graphs", PLoS ONE (2011). [3] M. Tumminello, T. Aste, T. Di Matteo, R. N. Mantegna, PNAS 102, n. 30 (2005) 10421.