Wrtpulse: Difference between revisions
Jump to navigation
Jump to search
imported>Neal |
imported>Neal |
||
| Line 9: | Line 9: | ||
===Description=== | ===Description=== | ||
This function rewrites vectors of system inputs and output so that they may be used with PLS and other modeling routines to obtain finite impulse response (FIR) and ARX models. If the inputs (u and y) are DataSet objects then so are the outputs (newu and newy). | |||
If no inputs are provided or an empty data matrix is provided, WRTPULSE will provide step-by-step loading of each input and the results will be written to the workspace. These results will include a model structure (settings) that can be used to reproduce the same decomposition on new data if passed in as (u) or (n). | |||
====Inputs==== | ====Inputs==== | ||
* '''u''' = a matrix of input vectors | * '''u''' = a M by N matrix of input vectors where the each input is a column vector. | ||
* '''y''' = | * '''y''' = the corresponding M by 1 output vector to be predicted by the model. | ||
* '''n''' = a row vector with the number of coefficents to use for each input | * '''n''' = a 1 by N row vector with the number of coefficents to use for each input (each element of '''n''' is the number of past periods to consider for each input). | ||
* '''delay''' = a row vector containing the number of time units of delay for each input | * '''delay''' = a 1 by N row vector containing the number of time units of delay for each input. | ||
====Outputs==== | ====Outputs==== | ||
Revision as of 11:16, 23 October 2013
Purpose
Create input/output matrices for dynamic model identification.
Synopsis
- [newu,newy] = wrtpulse(u,y,n,delay)
Description
This function rewrites vectors of system inputs and output so that they may be used with PLS and other modeling routines to obtain finite impulse response (FIR) and ARX models. If the inputs (u and y) are DataSet objects then so are the outputs (newu and newy).
If no inputs are provided or an empty data matrix is provided, WRTPULSE will provide step-by-step loading of each input and the results will be written to the workspace. These results will include a model structure (settings) that can be used to reproduce the same decomposition on new data if passed in as (u) or (n).
Inputs
- u = a M by N matrix of input vectors where the each input is a column vector.
- y = the corresponding M by 1 output vector to be predicted by the model.
- n = a 1 by N row vector with the number of coefficents to use for each input (each element of n is the number of past periods to consider for each input).
- delay = a 1 by N row vector containing the number of time units of delay for each input.
Outputs
- newu = a matrix of lagged input variables
- newy = the corresponding output vector