System Identification Toolbox

New Topic

1431 topics

Page 52 of 58
- Jump to page:
Previous
1
…
50
51
52
53
54
…
58
Next

Topics

Replies

Views

Last post

It is recommended to run model validation tests (RESID, COMPARE commands) using a dataset different from the one used fo

Last post by fyousefi « Tue Mar 02, 2021 7:01 am

by fyousefi » Tue Mar 02, 2021 7:01 am

0 Replies

642 Views

Last post by fyousefi
Tue Mar 02, 2021 7:01 am
There is a very strong indication that the dynamics of the model is not adequately described.

Last post by fyousefi « Tue Mar 02, 2021 7:00 am

by fyousefi » Tue Mar 02, 2021 7:00 am

0 Replies

578 Views

Last post by fyousefi
Tue Mar 02, 2021 7:00 am
There is a strong indication that the dynamics of the model is not adequately described.

Last post by fyousefi « Tue Mar 02, 2021 7:00 am

by fyousefi » Tue Mar 02, 2021 7:00 am

0 Replies

613 Views

Last post by fyousefi
Tue Mar 02, 2021 7:00 am
There is an indication that the dynamics of the model is not adequately described.

Last post by fyousefi « Tue Mar 02, 2021 6:59 am

by fyousefi » Tue Mar 02, 2021 6:59 am

0 Replies

631 Views

Last post by fyousefi
Tue Mar 02, 2021 6:59 am
A first general advice is to run RESID(VDATA FR') to check in which frequency ranges the model error is present. If

Last post by fyousefi « Tue Mar 02, 2021 6:58 am

by fyousefi » Tue Mar 02, 2021 6:58 am

0 Replies

674 Views

Last post by fyousefi
Tue Mar 02, 2021 6:58 am
In particular you should pay attention to lags from input { number,integer}. Modify KU and the model order so that

Last post by fyousefi « Tue Mar 02, 2021 6:57 am

by fyousefi » Tue Mar 02, 2021 6:57 am

0 Replies

594 Views

Last post by fyousefi
Tue Mar 02, 2021 6:57 am
There is no indication that the noise model should be incorrect.

Last post by fyousefi « Tue Mar 02, 2021 6:55 am

by fyousefi » Tue Mar 02, 2021 6:55 am

0 Replies

611 Views

Last post by fyousefi
Tue Mar 02, 2021 6:55 am
There is a very strong indication that the model errors are not white. Since you have not built a noise model, this is O

Last post by fyousefi « Tue Mar 02, 2021 6:55 am

by fyousefi » Tue Mar 02, 2021 6:55 am

0 Replies

640 Views

Last post by fyousefi
Tue Mar 02, 2021 6:55 am
There is an indication that the model errors are not white. Since you have not built a noise model, this is OK. However

Last post by fyousefi « Tue Mar 02, 2021 6:53 am

by fyousefi » Tue Mar 02, 2021 6:53 am

0 Replies

581 Views

Last post by fyousefi
Tue Mar 02, 2021 6:53 am
There is a very strong indication that the residuals are not white. To get a good noise model you need to increase the o

Last post by fyousefi « Tue Mar 02, 2021 6:53 am

by fyousefi » Tue Mar 02, 2021 6:53 am

0 Replies

613 Views

Last post by fyousefi
Tue Mar 02, 2021 6:53 am
There is a strong indication that the residuals are not white. To get a good noise model you need to increase the orders

Last post by fyousefi « Tue Mar 02, 2021 6:52 am

by fyousefi » Tue Mar 02, 2021 6:52 am

0 Replies

561 Views

Last post by fyousefi
Tue Mar 02, 2021 6:52 am
There is an indication that the residuals are not white. To get a good noise model you need to increase the orders assoc

Last post by fyousefi « Tue Mar 02, 2021 6:51 am

by fyousefi » Tue Mar 02, 2021 6:51 am

0 Replies

587 Views

Last post by fyousefi
Tue Mar 02, 2021 6:51 am
The model order seems unnecessarily high. Try models of lower order and see if they are able to reproduce validation dat

Last post by fyousefi « Tue Mar 02, 2021 6:50 am

by fyousefi » Tue Mar 02, 2021 6:50 am

0 Replies

514 Views

Last post by fyousefi
Tue Mar 02, 2021 6:50 am
On the other hand, the accuracy of the estimate seems to be such that the high order dynamics is accurately estimated.

Last post by fyousefi « Tue Mar 02, 2021 6:49 am

by fyousefi » Tue Mar 02, 2021 6:49 am

0 Replies

556 Views

Last post by fyousefi
Tue Mar 02, 2021 6:49 am
Using continuous-time initial condition object with discrete-time frequency-domain data can lead to inaccurate results.

Last post by fyousefi « Tue Mar 02, 2021 6:48 am

by fyousefi » Tue Mar 02, 2021 6:48 am

0 Replies

574 Views

Last post by fyousefi
Tue Mar 02, 2021 6:48 am
In the "initialCondition(SYS)" command, SYS must be a numeric, or parametric LTI object, such as IDSS, IDTF, IDPOLY, TF,

Last post by fyousefi « Tue Mar 02, 2021 6:47 am

by fyousefi » Tue Mar 02, 2021 6:47 am

0 Replies

534 Views

Last post by fyousefi
Tue Mar 02, 2021 6:47 am
The initialCondition object cannot be created by converting from the object.

Last post by fyousefi « Tue Mar 02, 2021 6:47 am

by fyousefi » Tue Mar 02, 2021 6:47 am

0 Replies

580 Views

Last post by fyousefi
Tue Mar 02, 2021 6:47 am
Model arrays cannot be used to create initialCondition object.

Last post by fyousefi « Tue Mar 02, 2021 6:46 am

by fyousefi » Tue Mar 02, 2021 6:46 am

0 Replies

575 Views

Last post by fyousefi
Tue Mar 02, 2021 6:46 am
Continuous-time models containing feedthrough cannot be used to create an initialCondition object.

Last post by fyousefi « Tue Mar 02, 2021 6:45 am

by fyousefi » Tue Mar 02, 2021 6:45 am

0 Replies

507 Views

Last post by fyousefi
Tue Mar 02, 2021 6:45 am
When initialCondition.X0 contains greater than 1 columns, the system SYS = idss(initialCondition) must be expressible as

Last post by fyousefi « Tue Mar 02, 2021 6:45 am

by fyousefi » Tue Mar 02, 2021 6:45 am

0 Replies

632 Views

Last post by fyousefi
Tue Mar 02, 2021 6:45 am
When comparing the response of a discrete-time model, the sample time of the model must match that of validation data

Last post by fyousefi « Tue Mar 02, 2021 6:34 am

by fyousefi » Tue Mar 02, 2021 6:34 am

0 Replies

587 Views

Last post by fyousefi
Tue Mar 02, 2021 6:34 am
The initial states and model predictors will be computed separately for each output because the model contains irreducib

Last post by fyousefi « Mon Mar 01, 2021 8:48 pm

by fyousefi » Mon Mar 01, 2021 8:48 pm

0 Replies

572 Views

Last post by fyousefi
Mon Mar 01, 2021 8:48 pm
Response values containing non-finite values were ignored for the computation of goodness of fit measures.

Last post by fyousefi « Mon Mar 01, 2021 8:47 pm

by fyousefi » Mon Mar 01, 2021 8:47 pm

0 Replies

544 Views

Last post by fyousefi
Mon Mar 01, 2021 8:47 pm
The output values for time instants less than those corresponding to the provided input signal will not be computed and

Last post by fyousefi « Mon Mar 01, 2021 8:47 pm

by fyousefi » Mon Mar 01, 2021 8:47 pm

0 Replies

563 Views

Last post by fyousefi
Mon Mar 01, 2021 8:47 pm
Provide a dataset for predicting or simulating model response.

Last post by fyousefi « Mon Mar 01, 2021 8:46 pm

by fyousefi » Mon Mar 01, 2021 8:46 pm

0 Replies

478 Views

Last post by fyousefi
Mon Mar 01, 2021 8:46 pm

New Topic

1431 topics

Page 52 of 58
- Jump to page:
Previous
1
…
50
51
52
53
54
…
58
Next

Return to Board Index

Who is online

Users browsing this forum: No registered users and 1 guest

Forum permissions

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum