On Stability of Oja Algorithm

Radosław Sikora , Władysław Skarbek

Abstract

By elementary tools of matrix analysis, we show that the discrete dynamical system defined by Oja algorithm is stable in the ball K(0,81/64) if only gains β n are bounded by (2B)−1, where B = b 2 and b is the bound for the learning sequence. We also define a general class of Oja’s systems (with gains satisfying stochastic convergence conditions) which tend to the infinity with exponential rate if only their initial states are chosen too far from the zero point.
Author Radosław Sikora RE
Radosław Sikora,,
- The Institute of Radioelectronics
, Władysław Skarbek RE
Władysław Skarbek,,
- The Institute of Radioelectronics
Pages354-360
Book Polkowski Lech, Skowron Andrzej (eds.): Rough Sets and Current Trends in Computing, Lecture Notes In Computer Science, vol. LNAI 1424, 1998, Springer, ISBN 3-540-64655-8, 625 p., DOI:10.1007/3-540-69115-4
Keywords in EnglishArtificial Intelligence (incl. Robotics), Computation by Abstract Devices, Image Processing and Computer Vision, Mathematical Logic and Formal Languages
URL http://link.springer.com/chapter/10.1007/3-540-69115-4_48
Languageen angielski
Score (nominal)7
Citation count*6 (2018-02-13)
Cite
Share Share



* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.
Back