Learning Systems : Stochastic Automata Models

  • M.A.L. Thathachar Indian Institute of Science, Banglore
Keywords: Threshold Asymptotically, Class conditional densities, Modelling learning systems

Abstract

Stochastic automata operating in an unknown random environment have been successfully used in modelling learning systems. For such learning automata a new algorithm which estimates the environmental characteristics is presented and is used in the solution of a pattern classification problem with unknown class-conditional densities. The algorithm selects the optimal threshold asymptotically with probability one.

Author Biography

M.A.L. Thathachar, Indian Institute of Science, Banglore
Indian Institute of Science, Banglore-560 012
Published
2014-01-24
How to Cite
ThathacharM. (2014). Learning Systems : Stochastic Automata Models. Defence Science Journal, 35(3), 361-366. https://doi.org/10.14429/dsj.35.6029
Section
Special Issue Papers