GKA Gaussian Kernel Algorithm Correlation Dimension Estimation Estimation of correlation dimension, entropy and noise level using the GKA algorithm. INSTALLATION - Place the contents of this archive in a directory on MATLAB's search path. - Compile the mex functions. On recent versions of MATLAB, >> mex interbinref.c >> mex interbin.c should be sufficient. Older versions may require you to run mex -setup first. If you know nothing about mex files, start by reading the help. - run test.m USAGE See individual help files and example in test.m DISTRIBUTION Please do not redistribute these programs without my permission. This distribution can be obtained directly from the author. If you use these programs for anything interesting, let me know. If you use these programs in a published work please cite the appropriate source(s). DISCLAIMER Estimating invariants from noisy time series can be risky. Do *not* trust the output of this algorithm blindly. Any liability or problems caused are yours! SOURCES C. Diks, "Estimating invariants of noisy attractors" Physical Review E 53 (1996): R4263-R4266 DJ Yu, M Small, R.G. Harrison and C. Diks, "Efficient implemntation of the Gaussian kernel algorithm in estimating invariants and noise level from noisy time series data" Physical Review E 61 (2000): 3750-3756 CONTACT For further information or problems, please contact the author: Michael Small Electronic and Information Engineering Hong Kong Polytechnic University Hung Hom, Kowloon, Hong Kong email: ensmall@polyu.edu.hk 25/2/2002