Data ranking and clustering via normalized graph cut based on asymmetric affinity

O. Kyrgyzov (1) , Isabelle Bloch (1) , Y. Yang (1) , J. Wiart (2) , A. Souloumiac (3)

1 WHIST Lab, Institut Mines Telecom, 46 rue Barrault 75634 Paris Cedex 13, France
2 Orange Labs R&D [Rennes]
3 LADIS - Laboratoire d'analyse des données et d'intelligence des systèmes

In this paper, we present an extension of the state-of-the-art normalized graph cut method based on asymmetry of the affinity matrix. We provide algorithms for classification and clustering problems and show how our method can improve solutions for unequal and overlapped data distributions. The proposed approaches are based on the theoretical relation between classification accuracy, mutual information and normalized graph cut. The first method requires a priori known class labeled data that can be utilized, e.g., for a calibration phase of a brain-computer interface (BCI). The second one is a hierarchical clustering method that does not involve any prior information on the dataset.

Keywords: asymmetric affinity; BCI; Classification accuracy; Classification and clustering; Data distribution; Graph cut; Hierarchical clustering methods; Mutual informations; Brain computer interface; Cluster analysis; Image analysis; Graphic methods

Domains: Numerical Analysis [math.NA]; Data Analysis, Statistics and Probability [physics.data-an]; General Topology [math.GN]; Signal and Image Processing

O. Kyrgyzov, Isabelle Bloch, Y. Yang, J. Wiart, A. Souloumiac. Data ranking and clustering via normalized graph cut based on asymmetric affinity. 17th International Conference on Image Analysis and Processing – ICIAP 2013, Sep 2013, Naples, Italy. pp.562-571, ⟨10.1007/978-3-642-41184-7_57⟩. ⟨cea-01831051⟩