Neural Network Toolbox

Index

A

ADALINE network

decision boundary <1> <2>

Adaption

custom function

Adaptive filter

noise cancellation example

prediction application

prediction example

Adaptive linear networks <1> <2>

Amplitude detection

Applications

adaptive filtering

medical <1> <2>

oil and gas exploration

telecommunications

Architecture

bias connection <1> <2>

input connection <1> <2>

layer connection <1> <2>

number of inputs <1> <2>

number of layers <1> <2>

number of outputs <1> <2>

number of targets <1> <2>

output connection <1> <2>

target connection <1> <2>

B

Backpropagation <1> <2>

Backtracking search

Batch training <1> <2> <3>

dynamic networks

static networks

Bayesian framework

benchmark <1> <2>

BFGS quasi-Newton algorithm

Bias

initialization function

learning function

learning parameters

subobject <1> <2>

C

Cell array

initial layer delay states

inputs and targets

inputs property

layers property

matrix of concurrent vectors

matrix of sequential vectors

sequence of outputs

sequential inputs

tap delayed inputs

weight matrices and bias vectors

Charalambous' search

Code

mathematical equivalents

perceptron network

Competitive layer

Competitive neural network

Competitive transfer function <1> <2> <3>

Concurrent inputs <1> <2>

Conjugate gradient algorithm

Fletcher-Reeves update

Polak-Ribiere update

Powell-Beale restarts

Continuous stirred tank reactor

Control

electromagnet <1> <2>

feedback linearization <1> <2>

model predictive control <1> <2> <3> <4> <5> <6>

model reference control <1> <2> <3> <4> <5> <6>

NARMA-L2 <1> <2> <3> <4> <5> <6>

plant <1> <2> <3>

Custom

D

Decision boundary <1> <2>

Demonstrations

nnd12sd1 <1> <2>

Distance <1> <2>

custom function

Dynamic network <1> <2>

Dynamic networks

training <1> <2>

E

Electromagnet <1> <2>

recurrent connection

Euclidean distance

networks <1> <2>

F

Feedback linearization <1> <2>

Feedforward network

Finite impulse response filter <1> <2>

Fletcher-Reeves update

G

Generalized regression network

Golden section search

Gradient descent algorithm <1> <2>

with momentum <1> <2>

Graphical User Interface

Gridtop topology

H

Hagan, Martin <1> <2>

Hard limit transfer function <1> <2> <3>

Heuristic techniques

Hidden layer

Hopfield network

design equilibrium point

solution trajectories

stable equilibrium point

target equilibrium points

Hybrid bisection-cubic search

I

networks <1> <2>

training data <1> <2>

Incremental training

Initial step size function

Initialization

additional functions

custom function

parameters <1> <2>

Input

subobject <1> <2> <3>

Input vector

dimension reduction

Input weights

Inputs

sequential <1> <2>

Installation

J

Jacobian matrix

K

Kohonen learning rule

L

Lambda parameter

Layer

distance function

initialization function

net input function

topology function

transfer function

Layer weights

Learning rate

Hebb with decay

supervised learning

unsupervised learning

Widrow-Hoff <1> <2> <3> <4> <5> <6>

Learning vector quantization

learning rule <1> <2>

union of two sub-classes

Least mean square error <1> <2>

Levenberg-Marquardt algorithm

Line search function

backtracking search

Charalambous' search

Golden section search

hybrid bisection-cubic search

Linear networks

Linear transfer function <1> <2> <3> <4> <5>

Linearly dependent vectors

Log-sigmoid transfer function <1> <2> <3>

M

Manhattan distance

Maximum performance increase

Maximum step size

Mean square error function

Memory reduction

Model predictive control <1> <2> <3> <4> <5> <6>

Model reference control <1> <2> <3> <4> <5> <6>

Momentum constant

N

NARMA-L2 <1> <2> <3> <4> <5> <6>

Net input function

Network

dynamic <1> <2>

Network Function

Network layer

Network/Data Manager window

Neural network

adaptive linear <1> <2>

generalized regression

multiple layer <1> <2> <3>

one layer <1> <2> <3> <4> <5>

self organizing

self-organizing feature map

Neural Network Design

Instructor's Manual

Neuron

dead (not allocated)

Newton's method

NN predictive control <1> <2> <3> <4> <5> <6>

Normalization

inputs and targets

mean and standard deviation

Notation

abbreviated <1> <2>

transfer function symbols <1> <2>

Numerical optimization

O

One step secant algorithm

Ordering phase learning rate

Outlier input vector

Output

subobject <1> <2>

Output layer

Overdetermined systems

P

Pass

pattern recognition

Perceptron learning rule <1> <2>

Perceptron network

Performance function

Plant <1> <2> <3>

Plant identification <1> <2> <3> <4>

Polak-Ribiere update

Post-training analysis

Powell-Beale restarts

Predictive control <1> <2> <3> <4> <5> <6>

Principal component analysis

Probabilistic neural network

Q

Quasi-Newton algorithm

R

Radial basis

efficient network

Radial basis transfer function

Recurrent connection

Recurrent networks

Resilient backpropagation

S

Self-organizing feature map (SOFM) network

one-dimensional example

two-dimensional example

Self-organizing networks

Sequential inputs <1> <2>

Sigma parameter

SIMULINK

generating networks

NNT block set <1> <2>

Spread constant

Squashing functions

Static networks

Subobject

bias <1> <2> <3>

input <1> <2> <3> <4>

output <1> <2> <3>

target <1> <2> <3>

weight <1> <2> <3> <4> <5>

Supervised learning

System identification <1> <2> <3> <4> <5> <6>

T

Tan-sigmoid transfer function

Tapped delay line <1> <2>

Target

subobject <1> <2>

Topologies

Topologies for SOFM neuron locations

custom function

competitive networks

custom function

definition <1> <2>

post-training analysis

self organizing feature map

Training styles

Training with noise

Transfer function

competitive <1> <2> <3>

hard limit <1> <2>

linear <1> <2> <3>

log-sigmoid <1> <2> <3>

saturating linear

triangular basis

Transformation matrix

Tuning phase learning rate

Tuning phase neighborhood distance

U

Underdetermined systems

Unsupervised learning

V

Variable learning rate algorithm

Vectors

linearly dependent

W

Weight

initialization function <1> <2>

learning <1> <2>

learning function <1> <2>

learning parameters <1> <2>

subobject <1> <2> <3>

value <1> <2> <3>

weight function <1> <2>

Weight function

Weight matrix

Widrow-Hoff learning rule <1> <2> <3> <4> <5> <6>

Workspace (command line)

Argument Checking