|
Point set of uniformity, the traditional method: crowding index, polyethylene mass index, dispersion index, information entropy and variance / mean. However, the foregoing method in addition to the information entropy, its essence is the variance to describe the degree of uniformity of the distribution of the data, is generally believed that the variance of the data or of sample points is smaller, the degree of dispersion is smaller, then the uniformity of the smaller. However, some studies on specific issues, there is a case: the variance smaller, evenness the contrary, the large (for a very uniform distribution of data, the evenness is great, but its variance is small). Information entropy can be accurately measured uniform, however, too computationally intensive, in the division of the space, there are also the scale selection problem. Respect to the method described above, uniformity not only overcome the variance on the uniform set of points can not characterize defects, and calculate fully dependent on the track system, the calculation amount is small. Paper core instantaneous chaotic intensity uniformity theory (ICM) and K-step transient chaotic intensity as a research tool, track rules, the quasi-periodic orbits, ICM sequence of periodic orbits and random orbit; attempt to measure the uniformity information In addition, given the numerical simulation, and synchronization information entropy better; ICM column two parameters determine criteria uniformity to resolve the securities market previous unrest and established the multi-objective decision-making model based on the uniformity of the securities market . The main results obtained by the paper are as follows: the ICM column characteristics (1) track 1 track X = {x 1 , x 2 , ..., the x N < / sub>} from the random system F, then any number of empty travel K2, the ICM are random sequence. Two pairs of periodic orbits X = {x , x 2 , ..., the x N }, when the number of actual travel K1 meet K < sub> 1 lt; T when the track long enough, where T is the orbital period of any given number of empty travel K2 have: ICM is a periodic orbit, and the cycle T 0 lt ; T. 3 aligned periodic orbit X = {x 1 x 2 , ..., the x N }, the ICM column is a series of quasi-periodic. 4 guesses: In addition to the above movement outside regular motion and chaotic motion, when the number of empty travel K2 is large enough, ICM sequence is a random sequence. 5 for random ICM sequence, whether independent identically distributed and the number of empty travel K2, K2 is greater than a constant K0 showed independent and identically distributed. Random ICM sequence distribution function for the normal distribution, and the distribution function curve track randomness enhanced more and more fat. (2) the ICM and Information Entropy synchronization through information entropy metric information and ICM resolve even the nature of the study, found that both have a lot of similarities, kent mapping and numerical simulation of the Lorenz attractor, proved ICM synchronization is very good information entropy. 2 to establish a multi-objective decision model based on uniformity: the i-securities transaction rates, i kind of securities investment yield, expected yield for the n securities investment portfolio. (3) parameters determined by two criteria for the selection of the number of empty travel K2 same time delay selection, there are a variety of methods, such as auto-correlation function method, the average displacement method and the average mutual information function method specifically identified time, can be based on the following two criteria: First, the phase space is extended so that the phase space trajectory extended from the main diagonal of the phase space as much as possible, but do not overlap; reduce the elements within each fragment, while maintaining each fragment contains the information of the dynamic system; the 2 real travel number K1, we define: E (k) = the ΔS CM = (K Δk) SCM? KSCM called kSCM change rate. K is defined: E * (k) = E (k) / Δk, the call kSCM average rate of change, i.e. the k-step instantaneous chaotic average rate of change of the strength of the take E * (k) tends to a constant when the real travel number. The simulation of a large number of measured data show that the stability of the guidelines.
|