Time-series data change point detecting method and a program for the future time-series data values ​​of the probability density distribution prediction method and program

摘要

the present invention, the true market price P (t) and center price PM (t ) and to apply the particulate filter method PUCK model to calculate the true market value P (t + 1) at time (t) determined by the. First, to obtain the probability density function of the parameters by generating a group of particles having a parameter representing the state of the PUCK model with different values. Then evaluating the fitness of each particle, and resample the particles as follows, depending on the goodness of fit. Generates a random number, compares the random number with a predetermined value, if the random number is greater than a predetermined value, the probability density function of the normal distribution such that the mean value of the parameter values of the model at time (t) according to regenerate the particles, if the random number is smaller than a predetermined value, to regenerate the particle with uniform distribution as a probability density function. We will continue these series of operations.