Hello All, Since it appears this community has people who have experience with a wide range of statistics and dynamics related concepts, i'll throw this out for someone to answer.
What is the correct embedding for an iterated cosine map specified as: X_t = lambda*cos(2pi*X_t-1) + eps; where eps is the usual white noise. I am using time-delayed mutual information and false nearest neighbors statistics to try to 'guess' the correct delay and embedding dimension for the attractor; but my answers don't jive with my intuition about what it 'should be' from a visual inspection of the above function. I would appreciate any insightful comments on this. Any pointers to quick-and-easy references would be great too! P ps. i'm not sure if its from this list, but i am getting empty "returned" mails to addresses i've never sent emails to; i suspect some of the users on this list no longer exist, but never bothered to leave the list officially. I am fairly sure i'm virus free given that our lab errs on the side of being hyer-freakish about security updates, etc. _____________________________________ Pradyumna Sribharga Upadrashta, PhD Student Scientific Computation, UofMN . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
