Code covered by the BSD License

Highlights from
The N-way Toolbox

B=unimodal(X,Y,Bold) UNIMODAL unimodal regression
B=unimodalcrossproducts(XtX,X... UNIMODALCROSSPRODUCTS
C=ckron(A,B) CKRON
G=T3core(X,Load,Weights,NonNe... T3CORE calculate Tucker core
G=neye(Fac); NEYE Produces a super-diagonal array
Options=setopts(Model); SETOPTS
XCorr=SetNaNs1(X,ExMin,ExMax,... SETNANS1 fluorescence artifact treatment
[A,B,C,D,E,F,G,H,I,J,K,L,M,N,... FAC2LET Convert 'Factors' to component matrices
[A,B,C,fit]=dtld(X,F,SmallMod... DTLD direct trilinear decomposition
[A,B,C]=gram(X1,X2,F); GRAM generalized rank annihilation method
[A]=cmatrep(A,crit,v1,v2) CMATREP optimized matrep
[Comb,ExplVariat]=explcore(G,... EXPLCORE For interpretation of cores and arrays
[Consistency,G,stdG,Target]=c... CORCOND Core consistency for PARAFAC model
[DIA_Rel,DIA_Abs]=CoreDian(C) COREDIAN Calculates core diagonality
[E]=GSM(V); GSM orthogonalization
[F,SS]=nonneg(X,M,F,Options); NONNEG1 alternative to NNLS
[Factors,G,ExplX,Xm]=tucker(X... TUCKER multi-way tucker model
[Factors,it,err,corcondia]=pa... PARAFAC multiway parafac model
[Factors]=ini(X,Fac,MthFl,Ign... INI initialization of loadings
[Factors]=inituck(X,Fac,MthFl... INITUCK initialization of loadings
[G]=calcore(X,Factors,Options... CALCORE Calculate the Tucker core
[Gv,ORot1,ORot2,ORot3]=maxdia... MAXDIA3 Maximize core diagonality
[Gv,ORot1,ORot2,ORot3]=maxswd... MAXSWD3 maximize core slice-wise diagonality
[Gv,ORot1,ORot2,ORot3]=maxvar... MAXVAR3 maximize core squared variance
[IndicesN]=Two2N(DimX,Indices... TWO2N Conversion of indices between unfoldings and N-way arrays
[R,Y]=complpol(X); COMPLPOL
[SWD_Rel,SWD_Abs]=CoreSWDn(C) CORESWDN Calculates the 'core-slice-wise-diagonality'.
[T,P]=fnipals(X,w,T) FNIPALS nipals algorithm for PCA
[Var_Rel,Var_Abs]=CoreVarn(C) COREVARN Calculates the 'core-variance'
[XValResult,Model]=ncrossreg(... NCROSSREG cross-validation of regression model
[X]=missmult(A,B) MISSMULT product of two matrices containing NaNs
[Xm]=nmodel(Factors,G,Om); NMODEL make model of data from loadings
[Xnew,mX,sX]=nprocess(X,Cent,... NPROCESS pre and postprocessing of multiway arrays
[Y,D]=normit(X) NORMIT normalize
[Y]=Refold3(X,R); REFOLD3
[]=plotfac(Factors,H,Axis0,Ax... PLOTFAC Plot the contents of Factors
[b,All,MaxML]=ulsr(x,NonNeg);
[b,B,AllBs]=monreg(x); MONREG monotone regression
[i,j]=getindxn(R,Idx); GETINDXN
[mm]=misssum(X,def) MISSSUM sum of a matrix X with NaN's
[v]=nsetdiff(A,B); NSETDIFF
[ypred,T,ssX,Xres]=npred(X,Fa... NPRED prediction with NPLS model
dC=derdia3(C,n) DERDIA3 the derivative of the nth mode of the core
dC=dervar3(C,n) DERVAR3 for core rotations
eemtimize(X,maxlv); EEMTIMIZE
fastnnls(XtX,Xty,tol) FASTNNLS Fast version of built-in NNLS
kr(A,B); KR Khatri-Rao product
krb(A,B); KRB Khatri-Rao product
load=pfls(ZtZ,ZtX,dimX,cons,O... PFLS
mm=missmean(X) MISSMEAN mean of a matrix X with NaN's
ncosine(factor1,factor2); NCOSINE multiple cosine/Tuckers congruence coefficient
ncrossdecomp(Method,X,FacMin,... NCROSSDECOMP crossvalidation of PARAFAC/Tucker/PCA
ncrossdecomp(Method,X,FacMin,... NCROSSDECOMP crossvalidation of PARAFAC/Tucker/PCA
nident(J,order,mag); NIDENT make 'identity' multi-way array
npls(X,Y,Fac,show); NPLS multilinear partial least squares regression
nshape.m
ntimes(X,Y,modeX,modeY); NTIMES Array multiplication
pfplot(X,Factors,Weights,Opti... PFPLOT plot parafac model
pftest(NumRep,X,Fac,Options,c... PFTEST find the number of PARAFAC components
st=stdnan(X); STDNAN estimate std with NaN's
tbxStruct=demos DEMOS Demo list for The N-way Toolbox for MATLAB
tucktest(varargin) TUCKTEST makes a scree-plot for Tucker using from [1 1 1] to Fac models
contents.m
derswd3.m
fnnls.m
parademo.m
ppp.m
tuckdemo.m
tucker2.m
View all files

4.30769

4.3 | 13 ratings Rate this file 54 Downloads (last 30 days) File Size: 5.2 MB File ID: #1088

The N-way Toolbox

by Rasmus Bro

22 Nov 2001 (Updated 20 Mar 2012)

Tools for fitting multi-way (tensor) models such as PARAFAC

| Watch this File

File Information

Description

The N-way Toolbox for MATLAB is the most advanced toolbox for fitting multi-way models. It is free and provides means for fitting PARAFAC, Tucker, N-PLS, GRAM and TLD. Further updates and data sets are available at http://www.models.life.ku.dk/

A related function is "parafac2", also available on the File Exchange.

MATLAB release

MATLAB 7.2 (R2006a)

Tags for This File
Everyone's Tags	gram, multiway, npls(2), parafac(2), probability, statistics, threeway, trilinear, tucker3
Tags I've Applied
Add New Tags	Please login to tag files.

Please login to add a comment or rating.

Comments and Ratings (15)
12 Apr 2013	Edoardo
30 Oct 2012	Charles Yang
20 Mar 2012	peter	wow, you are releasing this for everyone to use for free !
15 Mar 2012	Kunal Anand
05 Jan 2012	Drel
25 Dec 2008	USTC Lawrence
27 Nov 2006	alireza mohammadpour	complet
10 Nov 2006	shaohua hu	well done!
05 Apr 2006	Dr. Michael Griffiths	An excellent package, well laid out, a clearly explained tutorial section and clearly presented results on-screen.
28 Apr 2005	friend ass	good enough
24 May 2004	ROYA MIRZAJANI
23 Mar 2004	Matheus Freitas	Require high virtual memory
18 Feb 2004	Luis Mesa
28 Oct 2003	david stanley	Outstanding!
27 Aug 2003	Sangdon Lee	Excellent for PCA and PLS and N-way analysis.

Updates
07 Jan 2009	Bugs removed and files updated to version 7 of matlab
08 Jan 2009	Bugs fixed and files updated to current matlab version
08 Jan 2009	bugs fixed and files updated to current matlab version
05 Jul 2009	Deleted obsolete copyright notes in m-files
07 Jul 2009	Removed license information
20 Mar 2012	This is a minor update aiming at making the Nway toolbox work seamlessly with the newer MATLAB versions (version 3.20)

Highlights from The N-way Toolbox

The N-way Toolbox

Highlights from
The N-way Toolbox