lans.txt

模式识别工具包

 	lans		- Display general information about LANS toolbox
 	lans(1)		- Display detail information about each LANS function
 			- dump to a file 'lans.txt' in lans or working directory
 _______________________________________________________________________________
 
        LANS Pattern Recognition Toolbox
 
 	(MATLAB 5.0 or higher required)
 
 	(C) 2000.08.13 Kui-yu Chang
 	http://lans.ece.utexas.edu/~kuiyu
 
 	This program is free software; you can redistribute it and/or modify
 	it under the terms of the GNU General Public License as published by
 	the Free Software Foundation; either version 2 of the License, or
 	(at your option) any later version.
 
 	This program is distributed in the hope that it will be useful,
 	but WITHOUT ANY WARRANTY; without even the implied warranty of
 	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 	GNU General Public License for more details.
 
 	You should have received a copy of the GNU General Public License
 	along with this program; if not, write to the Free Software
 	Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA
 	or check
 			http://www.gnu.org/
 _______________________________________________________________________________
 This messaged was invoked by the 'lans' command
 
 	- Do a help on each sub-libraries for detailed help,
 	  e.g.
 		help lans-doc
 		help lans-fextract
 
 	- Some lans-functions, when invoked without any parameters, gives a demo
 	  e.g.
 	  	lans_group
 	  	lans_pps
 		lans_classifier
 
 	- Some (deprecated) functions may require the following toolboxes, which
 	  may be available from http://lans.ece.utexas.edu
 		NETLAB
 		GTM Toolbox
 ______________________________________________________________________________
 Currently available sub-libraries:

lans-data	- Sample LANS datasets

lans-old	- Old/Deprecated functions (will be removed in future versions)
	cscat		- Plot class labeled data points
	hplotv		- Plot an n-D vector in overwrite mode
	lans_cplotmd	- Plot d-dimensional labelled vector 
	lans_cscatline	- Plot class labeled data and regressed line
	wavesmoo	- Wavelet Smoothing
	check_km	- Check k-means (required by k-means)
	k-means		- K-means clustering
	gtm_rspg1	- Modified version of gtm_rspg1 from the GTM toolbox
	finduniq	- Deprecated function, see lans_finduniq
	rvget		- Get a single record from existing vertical data record postion
	rvmake		- Create empty vertical record
	rvput		- Put a single record into existing vertical data record postion
	rvshift		- Shift a single record in from end, displacing head of record
	wrap		- Wrap around (modulus)
	tallyc		- Tally and sort classes 
	group		- Gather/sort same class data and indexes them 
	vdist		- Compute L2 metric of 2 col vectors
	vdist2		- Compute squared L2 metric of 2 col vectors (faster)
	str2row		- Convert a 1-D string to numbers  
	tryerr		- Invoke error and halt if value evaluates to true
	kstr2num	- Str2num version compatible with MATCOM 3.0
	dupavg		- Average values dictated by repeating row 
	finddup		- Categorize consecutive repetitions in 1D string/numbers
	lanspara	- Provide alphabetical list of parameters
	paraget		- Extract '-para val' parameters from parameter string

	alias.m		- aliases to append to .cshrc
	lansparse.pl	- PERL script that adds input arguements to g_***.cxx
	sim.m		- template file for simulation
    lans-matcom	- Template files/scripts for MATCOM (acquired by Mathworks in 1999)

lans-visual	- Visualization routines
	lans_equalize	- Stretch histogram to fill all ranges
	lans_fit2page	- Compute subplot rows/cols given d-D data 
	lans_plotaxis	- Plot axes
	lans_plotcovars	- Plot 2-D covariance matrices outlines
	lans_plotgrad	- Plot gradient in multi-dimensional space
	lans_plotmd	- Plot d-dimensional vector 
	lans_plotproj	- Plot projections of points onto a curve
	lans_scatline	- Plot original points with regressed lines in d-D 
	lans_plotgrid	- Plot 2-D grid arranged D-dimensional vector 
	lans_plot	- Plot D-dimensional vector across multiple classes
	lans_plotseg	- Plot disjoint segments
	lans_color	- Generate N distinct RGB colors
	lans_symbol	- Generate small # of unique plot color/symbols 

lans-stats	- Statistical Routines
	lans_chicdf1	- Chi-square cdf (1-D)
	lans_chiicdf1	- Inverse Chi-square cdf (1-D)
	lans_center	- Center data at mean
	lans_scale	- Scale data to lie within specified range
	lans_chipdf1	- Chi-square pdf (1-D)
	lans_gammacdf1	- Gamma cdf (1-D)
	lans_gammaicdf1	- Inverse Gamma cdf (1-D)
	lans_likelihood	- Compute complete & average likelihood for mixture model
	lans_stand	- Standardize data to zero mean and unit variance
	lans_whiten	- Spherize (whiten) data to zero mean and unit covariance
	lans_basis	- Initialize basis/mixture structure
	lans_gausspdf	- Evaluates Gaussian pdf @ multiple vector points 
	lans_gausspdf1	- Evaluates Gaussian pdf (1-D) at multiple scalar points
	lans_pdf	- Compute multiple pdf w.r.t. multiple bases
	lans_gammapdf1	- Gamma pdf (1-D)
	lans_posterior	- Compute posterior responsibility matrix/log likelihood
	lans_gausscdf1	- Gaussian cdf (1-D)
	lasn_gaussicdf1	- Inverse Gaussian cdf (1-D)
	lans_gausstest	- Standard multivariate Gaussian test (d-D)
	lans_eigsort	- Sorted eigen-decomposition for MATLAB and compilables

lans-sim	- Automated simulation (and report generation)
	lans_ex1	- Run 1 experiment (repeated K times)
	lans_exfill	- Fill remaining cols of a run with stopped value
	lans_exinit	- Initialize experiment (ex) structure
	lans_expush	- Push data onto an ex structure 
	lans_exstats	- Compute statistics for an ex structure	
	lans_log	- Append log file
	lans_logmove	- Move log file to 'done' directory indicates completion
	lans_matmove	- Move mat file to 'done' directory
	lans_sim	- Run experiments with different parameters
	lans_siminit	- Initialize empty LANS simulation (lsim) structure
	lans_simload	- Plot results of lans_sim saved data
	lans_varinit	- Initialize LANS variable structure
	lans_partition	- Partition LANS data to train/validate/test sets

lans-regress	- Regression routines
	cspline1	- Cubic smoothing spline 1-D
	cspline1p	- Cubic smoothing spline 1-D PERIODIC
	cspline1pslow	- Periodic cubic smoothing spline 1-D (unaccelerated)
	cspline1slow	- Cubic smoothing spline 1-D (unaccelerated)
	findw		- Compute weights for llr 
	llr		- Local Linear Regression
	lwavg		- Local Weighted Average (w.r.t. abscissae)
	plegendre	- Legendre polynomial
	smooth		- Smooth data using a variety of methods
	spread		- Spread out knots/data to remove repetitions
	wlsqr		- Weighted polynomial Least Square Fit @ x(px)

lans-patrec	- Pattern recognition routines
	knn1		- Find nearest k neighbours in 1 dim, sorted indices
	lans_centroid	- Compute centroid of a 2-D image
	lans_gmoment	- Compute geometric moments of a 2-D image 
	lans_invariant	- Center and scale image to fixed mass
	lans_imgscale	- Scale and center a binary image w.r.t. bounding shape
	lans_zi2nm	- Convert index to order n and repetition m 
	lans_zmoment	- Compute Zernike moments of a 2-D image 
	lans_zmrecon	- Reconstruct image from Zernike Moments
	lans_znm2i	- Compute cumulative count of order n and repetition m 
	lans_zpoly	- Compute Zernike polynomial over unit circle
	lans_zpossible	- Compute repetitions for Zernike moments of order n
	lans_imgcrop	- Crop/Fill image
	lans_classifier	- General LANS classifier framework
	lans_cidx2class	- Recover class name from unique integer label
	lans_class2cidx	- Assign unique integer label to class (non-numeric)
	lans_confuse	- Compute confusion matrix from ouput/desired classes

    lans-pose	- Pose estimation routines (custom)
	lans_posefe	- Extract features from pose images
	lans_poseimg	- Load/display a scaled/thresholded pose image
	lans_poseobj	- Initialize empty LANS pose object (i.e. structure)
	lans_cart2pose	- Convert cartesion eye position to pose angles
	lans_pose2cart	- Convert pose angles to cartesion eye positions
	lans_posetune	- Callibrate poses w.r.t. trained manifold
	lans_rotate	- Rotate 2-D image by an arbitrary angle

 
lans-nn		- Unsupervised neural network routines
	bsol1		- Batch self-organizing line (1 epoch/iteration)
	bsolk		- Batch self-organizing line (k epochs)
	getnpara	- Get parameters from npara vector
	getsnpara	- Get parameters from snpara vector
	plotkmap	- Plot 2D KSOFM (dist colormap if |w| >2) 
	putnpara	- Put parameters into npara vector
	putsnpara	- Put parameters into snpara vector
	som1		- Online self-organizing map (1 iteration) 
	som1e		- Batch Self Organizing Map (1 epoch) 
	somk		- Online self-organizing map (k iterations) 
	somke		- Batch self-organizing map (k epochs) 

lans-io		- Input/Output routines
	extractlans	- Extract LANS API parameter from a string
	killadj		- Kill (merge) set of adjacent numbers into head of set
	killdos		- Kill char 13 in DOS linefeeds comprising chars 10+13
	lans_load	- Load LANS API dataset file
	loadpara	- Load parameter file
	paraset		- Set one '-para val' formatted parameter in string
	skiptill	- Skip commented lines until desired phrase is detected
	strfilter	- Perform various string filtering operations 
	stripremark	- Strip remarks from a line

lans-gendata	- Artificial data generation routines
	lans_genball	- Generate uniform y within a ball in R^D
	lans_gencircle	- Generate 2-D circle with indept. Gaussian noise
	lans_sphere	- Generates angular spaced points on a unit sphere (3-D)
	lans_grid	- Generate regularly spaced points on rectangular grid
	lans_nonlinear	- Generate selected examples of nonlinear points
	lans_pcgrid	- Generate regularly spaced points on PCA 1/2-D grid
	lans_shape	- Generate normal samples about a shape

lans-fextract	- Feature extraction routines
	ipca		- Reconstruction from Principal Components
	lans_pca	- Principal Component Analysis (batch) 
	lda		- Linear Discriminant Analysis (batch)
	ldacrit		- Linear Discriminant Analysis criterion evaluator
	ldaiter		- Linear Discriminant Analysis iterator for Feature Selection
	lans_proj2line	- Project points onto a piecewise linear curve

    lans-pps	- Probabilistic Principal Surface (D=1,2,3 or higher)
	lans_pps	- Probabilistic Principal Surface computation
	lans_pps1	- 1 iteration/epoch of Probabilistic Principal Surface
	lans_ppsclean	- Clear unnecessary elements for efficient PPS storage
	lans_ppsgem1	- Generalized EM algorithm iteration of PPS
	lans_ppsgtm1	- Generative Topographical Mapping approximation for PPS
	lans_ppsinit	- Probabilistic Principal Surface (PPS) initialization
	lans_ppsplot	- Plots PPS manifold
	lans_ppspost	- Compute PPS posterior responsibility matrix/likelihood
	lans_ppsproj	- Project test data onto PPS in data/latent space


    lans-pcurve	- Principal Curve routines (D=1, may need NETLAB)
	lans_1speed	- Make a principal curve unit speed
	lans_expect	- Principal curve local averaging E(f|lambda)
	lans_fx		- Compute linearly interpolated principal curve from knots 
	lans_gmmpcurve1	- Make a principal curve unit speed
	lans_gradient	- Compute the gradient of a principal curve
	lans_orient	- Orient covariances of a principal surface
	lans_pcurvek	- Principal Curve iteration
	lans_pcurve1	- Principal curve single iteration
	lans_porder	- reorder principal surface centers
	lans_projdist	- Find [mse,dist] from data to nearest projection
	lans_project	- Project data onto a principal curve
	lans_psurf	- Principal curve generation/initialization
	lans_psurf2mix	- Translate key psurf structures to NETLAB's mix structure

    lans-psurf	- Principal Surface routines (D=1,2 square, based on GTM)
	lans_asmoothness- Compute smoothness of curve using change in angles!
	lans_findML	- Compute M and L indices for simulation
	lans_ontriplane	- Check if points project ONTO a triangular plane
	lans_proj2sqr	- Project data onto a square patch in high-D space
	lans_proj2tri	- Project data onto a triangular patch in high-D space
	lans_psurf1	- 1 iteration of psurf (GTM) training
	lans_psurfgrad	- Compute q gradient vectors in data space R^D
	lans_psurfinit	- Principal surface initialization/creation (GTM based)
	lans_psurfpgrid	- Draw psurf grid manifold
	lans_psurfproj	- Project data (approx.) onto principal surface
	lans_smoothness	- Compute smoothness of curve using 2nd derivatives
	lans_sphere2idx	- convert longitude/latitude to absolute storage index

 
lans-doc	- Some documentation and scripts
	lans_template	- Invoke editor on LANS template file lans-doc/.template
	l		- Shortcut for 'dir' or 'ls'
	lans		- Display general information about LANS toolbox
	ll		- Shortcut for 'dir' or 'ls'

lans-common	- Common Routines
	a2of		- Analog to 1-of-c binary conversion (c=#classes)
	lans_cellsize	- Measure dimensionality of each cell component
	lans_assign	- Assign default values for unspecified/zero variables
	lans_boundidx	- Find upper, lower, and nearest bound indices
	lans_clip	- Clip values exceeding range
	lans_clipimg	- Clip img matrix regions into a vector 
	lans_default	- Set default values for unspecified fields
	lans_diversify	- Perturb repeating values to make it non-repetitive
	lans_fac	- Compute matrix factorial
	lans_findstable	- Find index when value has settled
	lans_glb	- Find greatest lower bound (infimum) in a 1D sequence
	lans_gramschmidt- Compute orthogonal vectors from orthonormal space
	lans_hashinc	- Increment an entry to hash table 
	lans_hashinit	- Initialize a numerical hash structure
	lans_iseven	- Check if number is even
	lans_isint	- Check if number is integer
	lans_killnan	- Replace NaN entries by specified number
	lans_lub	- Find least upper bound (supremum) in a 1D sequence
	lans_mahadist	- Mahalanobis distance from a set of vector means
	lans_matinv	- Matrix inverse lemma
	lans_replace	- Replace values in a matrix/vector
	lans_unitvec	- Make col vectors unit-length (Euclidean)
	lans_hashget	- Get an entry from hash table 
	lans_hashadd	- Add an entry to hash table 
	lans_hashnew	- Create a new empty 1-D hash table 
	lans_hashrm	- Delete/remove an entry from hash table
	lans_clone	- Clone matrix horizontally
	lans_findcell	- Find position of a cell in a cell
	lans_lin2md	- Convert linearly stored col vectors to multi-D array
	lans_md2lin	- Convert multi-D stored col vectors to linear array
	lans_dist	- Compute distances between 2 sets of column vectors
	lans_ininterval	- Check if a point is bounded by the interval
	lans_intercept	- Find intercept of piecewise linear curve
	lans_clone1	- Repeat each col vectors by k times
	lans_remove	- Remove an index of values from a col/row vector 
	lans_findidx	- Find indices of max/min values of each col
	lans_finduniq	- Remove CONSECUTIVE repetitions from sorted col vectors
	lans_group	- Group data based on outputs
	lans_hashset	- Modify an entry item in hash table 
	lans_paraget	- Get parameter value from string 
	lans_word2code	- Code a word/string into a vector

lans-gui	- Graphical User Interfaces
	lans_menuaxis	- Attach log/linear axis selection menu to figure