Next: 6.16 range_search
Up: 6. Mex-Function Reference
Previous: 6.14 nn_search
State space based prediction using nearest neighbors. The algorithms computes one or more
nearest neighbors to an initial state vector. The images of the nearest neighbors are
used to estimate to image of the initial state vector. The next iteration uses the previously
computed image as new initial state vector [145].
Syntax:
- x = predict(pointset, length, k, stepsize, mode)
Input arguments:
- pointset - a N by D double matrix
containing the coordinates of the point set, organized as
N points of dimension D
- length - number of iterations (length of prediction)
- k - number of nearest neighbors
- stepsize - prediction stepsize, usually one
- mode - (optional) method to estimate image of initial state vector
- 0 - direct prediction, no weight is applied to neighbors
- 1 - direct prediction, biquadratic weight is applied to neighbors
- 2 - integrated prediction, no weight is applied to neighbors
- 3 - integrated prediction, biquadratic weight is applied to neighbors
Output arguments:
- x - data set as double matrix, size length by D
Next: 6.16 range_search
Up: 6. Mex-Function Reference
Previous: 6.14 nn_search
Contents
Copyright © 1997-2009 DPI Göttingen