Jelölések

Szimbólumok

a

vektor

aT

transzponált vektor

a MathType@MTEF@5@5@+=feaagCart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqkLspw0le9v8qqaqFD0xXdHaVhbbf9v8qqaqFr0xc9pk0xbba9q8WqFfeaY=biLkVcLq=JHqpepeea0=as0Fb9pgeaYRXxe9vr0=vr0=vqpWqaaeaabiGaciaacaqabeaadaqaaqaaaOqaamaafmaabaGaaCyyaaGaayzcSlaawQa7aaaa@3D25@

az a vektor euklideszi normája, a vektor hossza

a MathType@MTEF@5@5@+=feaagCart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqkLspw0le9v8qqaqFD0xXdHaVhbbf9v8qqaqFr0xc9pk0xbba9q8WqFfeaY=biLkVcLq=JHqpepeea0=as0Fb9pgeaYRXxe9vr0=vr0=vqpWqaaeaabiGaciaacaqabeaadaqaaqaaaOqaamaafmaabaGaaCyyaaGaayzcSlaawQa7aaaa@3D25@ p

az a vektor p-normája

A

mátrix

A-1

inverz mátrix

A MathType@MTEF@5@5@+=feaagCart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaacbeGaa8xqamaaCaaaleqabaGaaiiiGaaaaaa@3792@

pszeudó-inverz (Moore-Penrose inverz) mátrix

| A | MathType@MTEF@5@5@+=feaagCart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqkLspw0le9v8qqaqFD0xXdHaVhbbf9v8qqaqFr0xc9pk0xbba9q8WqFfeaY=biLkVcLq=JHqpepeea0=as0Fb9pgeaYRXxe9vr0=vr0=vqpWqaaeaabiGaciaacaqabeaadaqaaqaaaOqaamaaemaabaGaaCyqaaGaay5bSlaawIa7aaaa@3D00@

az A mátrix determinánsa

LT(A)

az A mátrix alsó háromszögmátrixa

b

eltolás

C

autókovariancia mátrix

C(.)

kritérium függvény

CN

az N-dimenziós folytonos függvények halmaza

d

kívánt válasz

E

várható érték operátor

h

VC dimenzió

H

Hesse mátrix

I

egységmátrix

J(w)

elsődleges kritériumfüggvény SVM-nél

k

diszkrét időindex

K

kernel mátrix

K(x,z)

kernel függvény

L(w)

veszteségfüggvény, hibafüggvény

Lε(w)

ε érzéketlenségi sávval rendelkező veszteségfüggvény

L(w,b,α)

Lagrange függvény

L(.)

likelihood függvény

m

várható érték vektor

M

bázisfüggvények száma

N

a bemeneti minták dimenziója

p

valószínűség, valószínűség sűrűségfüggvény

p{ x|Θ } MathType@MTEF@5@5@+=feaagCart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqkLspw0le9v8qqaqFD0xXdHaVhbbf9v8qqaqFr0xc9pk0xbba9q8WqFfeaY=biLkVcLq=JHqpepeea0=as0Fb9pgeaYRXxe9vr0=vr0=vqpWqaaeaabiGaciaacaqabeaadaqaaqaaaOqaaiaadchadaGadaqaaiaadIhadaabbaqaaiaahI5aaiaawEa7aaGaay5Eaiaaw2haaaaa@3FEF@

feltételes valószínűség sűrűségfüggvény

P

mintapontok száma

P{ x|Θ } MathType@MTEF@5@5@+=feaagCart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqkLspw0le9v8qqaqFD0xXdHaVhbbf9v8qqaqFr0xc9pk0xbba9q8WqFfeaY=biLkVcLq=JHqpepeea0=as0Fb9pgeaYRXxe9vr0=vr0=vqpWqaaeaabiGaciaacaqabeaadaqaaqaaaOqaaiaadcfadaGadaqaaiaadIhadaabbaqaaiaahI5aaiaawEa7aaGaay5Eaiaaw2haaaaa@3FCF@

feltételes valószínűség

Q(α)

duális kritériumfüggvény osztályozós SVM-nél

Q(α,α’)

duális kritériumfüggvény regressziós SVM-nél

R

autókorrelációs mátrix

MathType@MTEF@5@5@+=feaagCart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaeSyhHekaaa@3745@

valós számok halmaza

R(w)

kockázat, mint a súlyvektor függvénye

N MathType@MTEF@5@5@+=feaagCart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaeSyhHe6aaWbaaSqabeaacaWGobaaaaaa@3845@

valós szám N-esek halmaza

tr(A)

az A mátrix nyoma

xX MathType@MTEF@5@5@+=feaagCart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaacbeGaa8hEaiabgIGioJqaaiaa+Hfaaaa@393C@

bemenet és a bemeneti tér

x ¯ MathType@MTEF@5@5@+=feaagCart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqiVCI8FfYJH8YrFfeuY=Hhbbf9v8qqaqFr0xc9pk0xbba9q8WqFfeaY=biLkVcLq=JHqpepeea0=as0Fb9pgeaYRXxe9vr0=vr0=vqpWqaaeaabaGaaiaacaqaaeaadaqaaqaaaOqaaiqadIhagaqeaaaa@36F9@

az x változó átlaga (súlyozott átlaga)

x MathType@MTEF@5@5@+=feaagCart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqiVCI8FfYJH8YrFfeuY=Hhbbf9v8qqaqFr0xc9pk0xbba9q8WqFfeaY=biLkVcLq=JHqpepeea0=as0Fb9pgeaYRXxe9vr0=vr0=vqpWqaaeaabaGaaiaacaqaaeaadaqaaqaaaOqaamaaamaabaGaamiEaaGaayzkJiaawQYiaaaa@38B1@

az x változó középértéke

yY MathType@MTEF@5@5@+=feaagCart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaacbeGaa8xEaiabgIGioJqaaiaa+Lfaaaa@393F@

kimenet és a kimeneti tér

y

egydimenziós kimenet

w

súlyvektor

Δw MathType@MTEF@5@5@+=feaagCart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaacbeGaa8hLdiaa=Dhaaaa@37EF@

a súlyvektor megváltozása, módosítása

w 0

eltolás

W

súlymátrix

O(.)

ordó jelölés

ξ

gyengítő változó

λ

sajátérték

μ

tanulási tényező

σ

szélesség paraméter

α

Lagrange multiplikátor

β

Lagrange multiplikátor

δ

általánosított hiba, érzékenység

ε

hiba

MathType@MTEF@5@5@+=feaagCart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaeyicI4maaa@3759@

eleme egy halmaznak szimbólum

MathType@MTEF@5@5@+=feaagCart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaGaeyOHI0maaa@37D6@

tartalmazás szimbólum

MathType@MTEF@5@5@+=feaagCart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqiVCI8FfYJH8YrFfeuY=Hhbbf9v8qqaqFr0xc9pk0xbba9q8WqFfeaY=biLkVcLq=JHqpepeea0=as0Fb9pgeaYRXxe9vr0=vr0=vqpWqaaeaabaGaaiaacaqaaeaadaqaaqaaaOqaaiablQIivbaa@3716@

únió szimbólum

MathType@MTEF@5@5@+=feaagCart1ev2aaatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqiVCI8FfYJH8YrFfeuY=Hhbbf9v8qqaqFr0xc9pk0xbba9q8WqFfeaY=biLkVcLq=JHqpepeea0=as0Fb9pgeaYRXxe9vr0=vr0=vqpWqaaeaabaGaaiaacaqaaeaadaqaaqaaaOqaaiablMIijbaa@3709@

metszet szimbólum

a MathType@MTEF@5@5@+=feaagCart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaaCWaaeaacaWGHbaacaGLUJVaayz+4daaaa@3BD0@

az a-nál nagyobb vagy azzal egyenlő legkisebb egész

a MathType@MTEF@5@5@+=feaagCart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaWaayWaaeaacaWGHbaacaGLWJVaay5+4daaaa@3BD3@

a egész része

MathType@MTEF@5@5@+=feaagCart1ev2aqatCvAUfeBSjuyZL2yd9gzLbvyNv2CaerbuLwBLnhiov2DGi1BTfMBaeXatLxBI9gBaerbd9wDYLwzYbItLDharqqtubsr4rNCHbGeaGqipu0Je9sqqrpepC0xbbL8F4rqqrFfpeea0xe9Lq=Jc9vqaqpepm0xbba9pwe9Q8fs0=yqaqpepae9pg0FirpepeKkFr0xfr=xfr=xb9adbaqaaeGaciGaaiaabeqaamaabaabaaGcbaacceGae83bIenaaa@375F@

gradiens operátor

δ(.,.)

Kronecker delta

Rövidítések

AIC

Akaike information criterion

APEX

Adaptive principal component extraction

BP

back-propagation

BPTT

back-propagation through time

CART

classification and regression tree

CMAC

cerebellar model articulation controller

DCT

discrete cosine transform

EIV

errors-in-variables

EM

expectation-maximization

ERM

empirical risk minimization

FIR

finite impulse response

FIR-MLP

finite impulse response multilayer perceptron

FPE

final prediction error

GHA

generalised Hebbian algorithm

HMOE

hiererchical mixture of experts

ICA

independent component analysis

IIR

infinite impulse response

KBANN

knowledge based artificial neural network

KKT

Karush-Kuhn-Tucker

KLT

Karhunen-Loève transform,

LMS

least mean squares

LS

least squares

LS-SVM

least squares support vector machine

LS2-SVM

LS - least squares support vector machine

LVQ

learning vector quantization

MDL

minimum description length

ML

maximum likelihood

MLP

multilayer perceptron

MOE

mixture of experts

MRAC

model reference adaptive control

MSE

mean squared error

NARX

nonlinear autoregressive with exogenous input

NARMAX

nonlinear autoregressive moving average with exogenous input

NFIR

nonlinear FIR

NIC

neural information criterion

NJB

nonlinear Jenkins-Box

NOE

nonlinear output error

OBD

optimal brain demage

OBS

optimal brain surgeon

OLS

orthogonal least squares

PCA

principal component analysis

RBF

radial basis function

RLS

recursive least squares

RREF

reduced row echelon form

RRKRR

reduced rank kernel ridge regression

RSVM

reduced support vector machine

RTRL

real time recursive learning

SLT

statistical learning theory

SMO

sequential minimal optimization

SOM

self-organizing map

SRM

structural risk minimization

SSVM

smooth support vector machine

SVM

support vector machine

TDNN

time delay neural network

TSP

traveling salesman problem

VC

Vapnik-Chervonenkis

WMSVM

weighted margin support vector machine