Diffusion Strategies For In-Network Principal Component Analysis

Nisrine Ghadban (1, 2) , Paul Honeine (1) , Farah Mourad-Chehade (1) , Clovis Francis (2) , Joumana Farah (3)

1 LM2S - Laboratoire Modélisation et Sûreté des Systèmes
2 LU / ULB - الجامعة اللبنانية [بيروت] = Lebanese University [Beirut] = Université libanaise [Beyrouth]
3 Holy-Spirit University of Kaslik [Jounieh]

This paper deals with the principal component analysis in networks, where it is improper to compute the sample covariance matrix. To this end, we derive several in-network strategies to estimate the principal axes, including noncooperative and cooperative (diffusion-based) strategies. The performance of the proposed strategies is illustrated on diverse applications, including image processing and dimensionality reduction of time series in wireless sensor networks.

Mots clés: unsupervised learning; in-network principal component analysis; cooperative diffusion-based strategy; covariance matrix; Principal component analysis; network; Convergence; Eigenvalues and eigenfunctions; Time series analysis; Wireless sensor networks; Cost function; Index Terms-Principal component analysis; distributed processing; adaptive learning; Covariance matrices

Domaines: Apprentissage [cs.LG]; Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]; Vision par ordinateur et reconnaissance de formes [cs.CV]

Nisrine Ghadban, Paul Honeine, Farah Mourad-Chehade, Clovis Francis, Joumana Farah. Diffusion Strategies For In-Network Principal Component Analysis. Proc. 24th IEEE workshop on Machine Learning for Signal Processing (MLSP), 2014, Reims, France. pp.1 - 6, ⟨10.1109/MLSP.2014.6958849⟩. ⟨hal-01965993⟩