yanxg / nlsa Goto Github PK

NLSA ALGORITHMS FOR MATLAB

This repository contains a collection of Matlab classes implementing nonlinear 
Laplacian spectral analysis (NLSA) and related kernel algorithms for analyzing
time series generated by dynamical systems. 

The distribution contains the following directories:

examples/torus:
   Example scripts for data generated by a dynamical system on the torus

nlsa/3rd_party
   Third-party functions 

nlsa/classes
   NLSA classes and methods

nlsa/utils
   Additional supporting functions

To install the package, put the nlsa directory in your Matlab path. This
can be done by executing the command

>> addpath(genpath('nlsa'))

from the root directory of the distribution, or by launching Matlab from a 
directory containing the startup.m file provided. 

Basic documentation is provided in the class definition files:

   classes/@nlsaModel_base/nlsaModel_base.m
   classes/@nlsaModel/nlsaModel.m
   classes/@nlsaModel_den/nlsaModel_den.m
   classes/@nlsaModel_ose/nlsaModel_ose.m

See also the README file in examples/torus.

Disclaimer: This code is provided as an academic research tool with no 
guarantees about stability, performance, and correctness of numerical results.
For comments and to report bugs email [email protected].  
  
Main contributor: Dimitris Giannakis (NYU)

Research funded through NSF grant DMS-1521775, ONR grant N00014-14-1-0150, 
ONR MURI grant 25-74200-F7112, and ONR YIP grant N00014-16-1-2649.