Training of Recurrent Neural Network

Training of Recurrent Neural Network - Training of...

Info iconThis preview shows pages 1–8. Sign up to view the full content.

View Full Document Right Arrow Icon
Training of Recurrent Neural Training of Recurrent Neural Network Network ME 697Y Sidharth Balasubramaniam
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
A Hybrid of Genetic Algorithm and Particle A Hybrid of Genetic Algorithm and Particle Swarm Optimization (HGAPSO) Swarm Optimization (HGAPSO) y Structural/Functional mapping problem y GA’s are Unsatisfactory for complex problems, and many possible solutions exist for the RNN training. y PSO’s conducts searches in the population. ¾ Recurrent Networks Fully Connected Recurrent Neural Network (FCRNN)
Background image of page 2
TSK TSK - - Type Recurrent Fuzzy Network Type Recurrent Fuzzy Network (TRFN) (TRFN) 6. Defuzzification 5.Context Nodes- Defuzzification 4.Consequent Nodes 3.Rule Node 2.Membership Functions 1.Input Nodes
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
y Function definition and the calculation for TSK- TRFN y Gaussian Membership function is used. y Each recurrent fuzzy if-then rule in TRFN, consisting of rules and external inputs
Background image of page 4
HGAPSO HGAPSO
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
HGAPSO HGAPSO 1. Coding : Floating Point Coding 2. Initialization : Both GA and PSO work on the same population ‘Ps’ individuals. Initialization involves assigning all the initial parameters such as ‘N’, Neural nodes, ‘Nr’, Number of Neural Rules, Learning parameters in PSO, Mutation Probability are assigned. 3. Enhancement : Based on the fitness values the top half individuals are marked and are called ‘Elites’. These elites are enhanced before reproducing. Combination of Cognition only model and Social Only model.
Background image of page 6
HGAPSO HGAPSO
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Image of page 8
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 12/29/2011 for the course ME 680 taught by Professor Na during the Fall '10 term at Purdue University-West Lafayette.

Page1 / 17

Training of Recurrent Neural Network - Training of...

This preview shows document pages 1 - 8. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online