Undefined function 'predict' for input arguments of type 'classreg.learning.partition.RegressionPartitionedLinear'.

Question

Akira Kato 2020-8-9

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/577312-undefined-function-predict-for-input-arguments-of-type-classreg-learning-partition-regressionpart

评论： Akira Kato 2020-8-11

采纳的回答： Image Analyst

After training model with cross validation:

model{i}=fitrlinear(xtrain',ytrain(i,:)','Crossval','on');

I try to use predict function

predict(model{i},xtest);

This worked fine when i didn't have cross validation flags in the fitrlinear function because the output was linear regression model. I looked on wiki for RegressionPartionedLinear object and its function is also called predict, but matlab giving me "undefined function" error.

How can I use the cross validated model?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Image Analyst 2020-8-10

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/577312-undefined-function-predict-for-input-arguments-of-type-classreg-learning-partition-regressionpart#answer_477604

Toolbox_License_Info.m

What does this say

>> which -all predict

Also if you run the attached program, what does it say?

3 个评论
显示 1更早的评论隐藏 1更早的评论

Akira Kato 2020-8-10

I get this by running which -all predict:

C:\Program Files\MATLAB\R2019a\toolbox\stats\classreg\RegressionLinear.m % RegressionLinear method

C:\Program Files\MATLAB\R2019a\toolbox\stats\classreg\+classreg\+learning\+regr\RegressionModel.m % classreg.learning.regr.RegressionModel method

C:\Program Files\MATLAB\R2019a\toolbox\stats\classreg\+classreg\+learning\Predictor.m % classreg.learning.Predictor method

C:\Program Files\MATLAB\R2019a\toolbox\nnet\cnn\@DAGNetwork\predict.m % DAGNetwork method

And running the attached file gives me:

v =

1×13 struct array with fields:

Name

Version

Release

Date

Checking on Control System Toolbox...

"Control System Toolbox" is "Control_Toolbox"

You currently have a license to use Control System Toolbox.

Checking on DSP System Toolbox...

"DSP System Toolbox" is "Signal_Blocks"

You currently have a license to use DSP System Toolbox.

Checking on Data Acquisition Toolbox...

"Data Acquisition Toolbox" is "Data_Acq_Toolbox"

You currently have a license to use Data Acquisition Toolbox.

Checking on Deep Learning Toolbox...

"Deep Learning Toolbox" is "Neural_Network_Toolbox"

You currently have a license to use Deep Learning Toolbox.

Checking on Image Processing Toolbox...

"Image Processing Toolbox" is "Image_Toolbox"

You currently have a license to use Image Processing Toolbox.

Checking on Instrument Control Toolbox...

"Instrument Control Toolbox" is "Instr_Control_Toolbox"

You currently have a license to use Instrument Control Toolbox.

Checking on MATLAB...

"MATLAB" is "MATLAB"

You currently have a license to use MATLAB.

Checking on Optimization Toolbox...

"Optimization Toolbox" is "Optimization_Toolbox"

You currently have a license to use Optimization Toolbox.

Checking on Signal Processing Toolbox...

"Signal Processing Toolbox" is "Signal_Toolbox"

You currently have a license to use Signal Processing Toolbox.

Checking on Simulink...

"Simulink" is "SIMULINK"

You currently have a license to use Simulink.

Checking on Simulink Control Design...

"Simulink Control Design" is "Simulink_Control_Design"

You currently have a license to use Simulink Control Design.

Checking on Statistics and Machine Learning Toolbox...

"Statistics and Machine Learning Toolbox" is "Statistics_Toolbox"

You currently have a license to use Statistics and Machine Learning Toolbox.

Checking on Symbolic Math Toolbox...

"Symbolic Math Toolbox" is "Symbolic_Toolbox"

You currently have a license to use Symbolic Math Toolbox.

Done running Toolbox_License_Info.m.

Image Analyst 2020-8-11

在 MATLAB Online 中打开

I think that looks okay. So we can debug further, attach xtrain, and ytrain in a .mat file

save('answers.mat', 'xtrain', 'ytrain');

and your script.

Akira Kato 2020-8-11

I tried it again and it didn't throw me error, the models were not objects RegressionPartionedLinear but just Regression Linear objects for some reason. I honestly don't know what the difference between yesterday and today is, maybe RegressionPartionedLinear is intermediate object? idk. Thanks for your help and patience. If there's anything that comes up again with the same problem ill just continue this thread.

请先登录，再进行评论。