Error when using svmpredict in matlab

[SeaBroNoMeal]

I succeeded in creating a svm struct model with svmtrain in matlab, but when I tried to use svmpredict to do further prediction, only to find that the output of the svmpredict were all empty. What should I do? Below is the code I used to do the prediction. [predict_label_1,accuracy_1] = svmpredict_01(train_label,Train_matrix,model);

Thank you very much.

[SeaBroNoMeal]

I succeeded in creating a svm struct model with svmtrain in matlab, but when I tried to use svmpredict to do further prediction, only to find that the output of the svmpredict were all empty. What should I do? Below is the code I used to do the prediction. [predict_label_1,accuracy_1] = svmpredict_01(train_label,Train_matrix,model);

Thank you very much.

Sorry I forgot to mention that since the matlab version I used is the latest, and in matlab 'svmtrain' is also a function. and this function has been removed, but in the libsvm you provided 'svmtrain' is still used, so I changed the name of the mex file from 'svmtrain' to 'svmtrain_01' (Because if not doing so, matlab will automatically output an error message which said 'function 'svmtrain' has been removed, use 'fitcsvm' instead'), and in order to keep consistence, I also changed the 'svmpredict.mexw64' file to 'svmpredict_01.mexw64'. And this is why I used 'svmpredict_01' in the code above.

[cjlin1]

Could you show more details (e.g. training and screen output of traing)? SeaBroNoMeal writes:

I succeeded in creating a svm struct model with svmtrain in
matlab, but when I tried to use svmpredict to do further
prediction, only to find that the output of the svmpredict
were all empty. What should I do?
Below is the code I used to do the prediction.
[predict_label_1,accuracy_1] = svmpredict_01
(train_label,Train_matrix,model);

Thank you very much.

Sorry I forgot to mention that since the matlab version I used is the latest, and in matlab 'svmtrain' is also a function. and this function has been removed, but in the libsvm you provided 'svmtrain' is still used, so I changed the name of the mex file from 'svmtrain' to 'svmtrain_01' (Because if not doing so, matlab will automatically output an error message which said 'function 'svmtrain' has been removed, use 'fitcsvm' instead'), and in order to keep consistence, I also changed the 'svmpredict.mexw64' file to 'svmpredict_01.mexw64'. And this is why I used 'svmpredict_01' in the code above.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.*

[SeaBroNoMeal]

Could you show more details (e.g. training and screen output of traing)? SeaBroNoMeal writes: I succeeded in creating a svm struct model with svmtrain in matlab, but when I tried to use svmpredict to do further prediction, only to find that the output of the svmpredict were all empty. What should I do? Below is the code I used to do the prediction. [predict_label_1,accuracy_1] = svmpredict_01 (train_label,Train_matrix,model); Thank you very much. Sorry I forgot to mention that since the matlab version I used is the latest, and in matlab 'svmtrain' is also a function. and this function has been removed, but in the libsvm you provided 'svmtrain' is still used, so I changed the name of the mex file from 'svmtrain' to 'svmtrain_01' (Because if not doing so, matlab will automatically output an error message which said 'function 'svmtrain' has been removed, use 'fitcsvm' instead'), and in order to keep consistence, I also changed the 'svmpredict.mexw64' file to 'svmpredict_01.mexw64'. And this is why I used 'svmpredict_01' in the code above. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.*

I used the following test code to train a svm model, and the returned model is shown as follows. load carsmall; X = [Horsepower,Weight]; Y = MPG; n = randperm(size(X,1)); Train_data = X(n(1:80),:); Train_label = Y(n(1:80),:); Test_data = X(n(81:end),:); Test_label = Y(n(81:end),:); model_test = svmtrain_01(Train_label,Train_data); [predict_label,acc] = svmpredict_01(Test_label,Test_data,model_test);

And the returned value of svmpredict function were all empty.

[cjlin1]

what's the screen output of model_test = svmtrain_01(Train_label,Train_data); SeaBroNoMeal writes:

Could you show more details (e.g. training and screen
output of traing)? SeaBroNoMeal writes: I succeeded in
creating a svm struct model with svmtrain in matlab, but
when I tried to use svmpredict to do further prediction,
only to find that the output of the svmpredict were all
empty. What should I do? Below is the code I used to do the
prediction. [predict_label_1,accuracy_1] = svmpredict_01
(train_label,Train_matrix,model); Thank you very much.
Sorry I forgot to mention that since the matlab version I
used is the latest, and in matlab 'svmtrain' is also a
function. and this function has been removed, but in the
libsvm you provided 'svmtrain' is still used, so I changed
the name of the mex file from 'svmtrain' to 'svmtrain_01'
(Because if not doing so, matlab will automatically output
an error message which said 'function 'svmtrain' has been
removed, use 'fitcsvm' instead'), and in order to keep
consistence, I also changed the 'svmpredict.mexw64' file to
'svmpredict_01.mexw64'. And this is why I used
'svmpredict_01' in the code above. — You are receiving this
because you are subscribed to this thread. Reply to this
email directly, view it on GitHub, or mute the thread.*

I used the following test code to train a svm model, and the returned model is shown as follows. load carsmall; X = [Horsepower,Weight]; Y = MPG; n = randperm (size(X,1)); Train_data = X(n(1:80),:); Train_label = Y(n (1:80),:); Test_data = X(n(81:end),:); Test_label = Y(n (81:end),:); model_test = svmtrain_01(Train_label,Train_data); [predict_label,acc] = svmpredict_01 (Test_label,Test_data,model_test);

pj b do5

And the returned value of svmpredict function were all empty. fxb j_el

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.*

[SeaBroNoMeal]

what's the screen output of model_test = svmtrain_01(Train_label,Train_data); SeaBroNoMeal writes: Could you show more details (e.g. training and screen output of traing)? SeaBroNoMeal writes: I succeeded in creating a svm struct model with svmtrain in matlab, but when I tried to use svmpredict to do further prediction, only to find that the output of the svmpredict were all empty. What should I do? Below is the code I used to do the prediction. [predict_label_1,accuracy_1] = svmpredict_01 (train_label,Train_matrix,model); Thank you very much. Sorry I forgot to mention that since the matlab version I used is the latest, and in matlab 'svmtrain' is also a function. and this function has been removed, but in the libsvm you provided 'svmtrain' is still used, so I changed the name of the mex file from 'svmtrain' to 'svmtrain_01' (Because if not doing so, matlab will automatically output an error message which said 'function 'svmtrain' has been removed, use 'fitcsvm' instead'), and in order to keep consistence, I also changed the 'svmpredict.mexw64' file to 'svmpredict_01.mexw64'. And this is why I used 'svmpredict_01' in the code above. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.* I used the following test code to train a svm model, and the returned model is shown as follows. load carsmall; X = [Horsepower,Weight]; Y = MPG; n = randperm (size(X,1)); Train_data = X(n(1:80),:); Train_label = Y(n (1:80),:); Test_data = X(n(81:end),:); Test_label = Y(n (81:end),:); model_test = svmtrain_01(Train_label,Train_data); [predict_label,acc] = svmpredict_01 (Test_label,Test_data,model_test); pj b do5 And the returned value of svmpredict function were all empty. fxb j_el — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.*

It's a long screen output of training status and the last status is:

[cjlin1]

After training can you type

model_test

to see how the model looks like

SeaBroNoMeal writes:

what's the screen output of model_test = svmtrain_01
(Train_label,Train_data); SeaBroNoMeal writes: Could you
show more details (e.g. training and screen output of
traing)? SeaBroNoMeal writes: I succeeded in creating a svm
struct model with svmtrain in matlab, but when I tried to
use svmpredict to do further prediction, only to find that
the output of the svmpredict were all empty. What should I
do? Below is the code I used to do the prediction.
[predict_label_1,accuracy_1] = svmpredict_01
(train_label,Train_matrix,model); Thank you very much.
Sorry I forgot to mention that since the matlab version I
used is the latest, and in matlab 'svmtrain' is also a
function. and this function has been removed, but in the
libsvm you provided 'svmtrain' is still used, so I changed
the name of the mex file from 'svmtrain' to 'svmtrain_01'
(Because if not doing so, matlab will automatically output
an error message which said 'function 'svmtrain' has been
removed, use 'fitcsvm' instead'), and in order to keep
consistence, I also changed the 'svmpredict.mexw64' file to
'svmpredict_01.mexw64'. And this is why I used
'svmpredict_01' in the code above. — You are receiving this
because you are subscribed to this thread. Reply to this
email directly, view it on GitHub, or mute the thread.* I
used the following test code to train a svm model, and the
returned model is shown as follows. load carsmall; X =
[Horsepower,Weight]; Y = MPG; n = randperm (size(X,1));
Train_data = X(n(1:80),:); Train_label = Y(n (1:80),:);
Test_data = X(n(81:end),:); Test_label = Y(n (81:end),:);
model_test = svmtrain_01(Train_label,Train_data);
[predict_label,acc] = svmpredict_01
(Test_label,Test_data,model_test); pj b do5 And the
returned value of svmpredict function were all empty. fxb
j_el — You are receiving this because you commented. Reply
to this email directly, view it on GitHub, or mute the
thread.*

It's a long screen output of training status and the last status is: m7y2i6 a

— You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.*

[SeaBroNoMeal]

I have provided the property of the returning model above, it's:

After training can you type

model_test to see how the model looks like SeaBroNoMeal writes: what's the screen output of model_test = svmtrain_01 (Train_label,Train_data); SeaBroNoMeal writes: Could you show more details (e.g. training and screen output of traing)? SeaBroNoMeal writes: I succeeded in creating a svm struct model with svmtrain in matlab, but when I tried to use svmpredict to do further prediction, only to find that the output of the svmpredict were all empty. What should I do? Below is the code I used to do the prediction. [predict_label_1,accuracy_1] = svmpredict_01 (train_label,Train_matrix,model); Thank you very much. Sorry I forgot to mention that since the matlab version I used is the latest, and in matlab 'svmtrain' is also a function. and this function has been removed, but in the libsvm you provided 'svmtrain' is still used, so I changed the name of the mex file from 'svmtrain' to 'svmtrain_01' (Because if not doing so, matlab will automatically output an error message which said 'function 'svmtrain' has been removed, use 'fitcsvm' instead'), and in order to keep consistence, I also changed the 'svmpredict.mexw64' file to 'svmpredict_01.mexw64'. And this is why I used 'svmpredict_01' in the code above. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.* I used the following test code to train a svm model, and the returned model is shown as follows. load carsmall; X = [Horsepower,Weight]; Y = MPG; n = randperm (size(X,1)); Train_data = X(n(1:80),:); Train_label = Y(n (1:80),:); Test_data = X(n(81:end),:); Test_label = Y(n (81:end),:); model_test = svmtrain_01(Train_label,Train_data); [predict_label,acc] = svmpredict_01 (Test_label,Test_data,model_test); pj b do5 And the returned value of svmpredict function were all empty. fxb j_el — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.* It's a long screen output of training status and the last status is: m7y2i6 a — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.*

[SeaBroNoMeal]

Hello, are you still there ?

[laxmi-priya]

hello, can you find any solution? i am also getting the same error. thanks

[cjlin1]

Do you also get empty output of svmpredict?

If so, could you send us your input of calling svmpredict

On 2019-07-19 02:36, laxmi-priya wrote:

hello, can you find any solution? i am also getting the same error. thanks

-- You are receiving this because you commented. Reply to this email directly, view it on GitHub [1], or mute the thread [2]. [ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/cjlin1/libsvm/issues/132?email_source=notifications\u0026email_token=ABI3BHUJMBUR2ZDRCPMYY5LQAGDLDA5CNFSM4GOIHCFKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2LEJFA#issuecomment-513164436", "url": "https://github.com/cjlin1/libsvm/issues/132?email_source=notifications\u0026email_token=ABI3BHUJMBUR2ZDRCPMYY5LQAGDLDA5CNFSM4GOIHCFKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2LEJFA#issuecomment-513164436", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

Links:

[1] https://github.com/cjlin1/libsvm/issues/132?email_source=notifications&email_token=ABI3BHUJMBUR2ZDRCPMYY5LQAGDLDA5CNFSM4GOIHCFKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2LEJFA#issuecomment-513164436 [2] https://github.com/notifications/unsubscribe-auth/ABI3BHUOY2B7TRTNGHEE7HTQAGDLDANCNFSM4GOIHCFA

[laxmi-priya]

thanks ..the calling function is [predict_label] = svmpredict(1, fe, model, '-b 0'); fe=[16.6355703299288 5.50239876948526 0.129682384057353 1.46400155085363 4.09956767886523 8.53661824811379 3.12358725188358 -0.255924946091940 1.68789128840180 1.33213123121757 -1.05335395199421 0.486931074011915 0.723009120298510 -1.11653922246564 0.245289861986370 0.330718833003371 0.336561455519822 -0.448193560013909 0.0928036970400240 0.115427375679267 0.710506841834280 0.899065229298500 0.982948775277088] model=

[cjlin1]

Your model should have a field sv_indices But it's missing?? How did you do the training?

Then in prediction fields related to prob outputs On 2019-07-21 06:00, laxmi-priya wrote:

thanks ..the calling function is [predict_label] = svmpredict(1, fe, model, '-b 0'); fe=[16.6355703299288 5.50239876948526 0.129682384057353 1.46400155085363 4.09956767886523 8.53661824811379 3.12358725188358 -0.255924946091940 1.68789128840180 1.33213123121757 -1.05335395199421 0.486931074011915 0.723009120298510 -1.11653922246564 0.245289861986370 0.330718833003371 0.336561455519822 -0.448193560013909 0.0928036970400240 0.115427375679267 0.710506841834280 0.899065229298500 0.982948775277088] model= [1]

-- You are receiving this because you commented. Reply to this email directly, view it on GitHub [2], or mute the thread [3]. [ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/cjlin1/libsvm/issues/132?email_source=notifications\u0026email_token=ABI3BHXJ5IU7YY3XGPU6HKDQARMXRA5CNFSM4GOIHCFKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2ODAII#issuecomment-513552417", "url": "https://github.com/cjlin1/libsvm/issues/132?email_source=notifications\u0026email_token=ABI3BHXJ5IU7YY3XGPU6HKDQARMXRA5CNFSM4GOIHCFKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2ODAII#issuecomment-513552417", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

Links:

[1] https://user-images.githubusercontent.com/27006197/61591485-0ae2b180-abe5-11e9-80a1-de655b06a7f9.png [2] https://github.com/cjlin1/libsvm/issues/132?email_source=notifications&email_token=ABI3BHXJ5IU7YY3XGPU6HKDQARMXRA5CNFSM4GOIHCFKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD2ODAII#issuecomment-513552417 [3] https://github.com/notifications/unsubscribe-auth/ABI3BHXKZKCVTI2BGHMSEELQARMXRANCNFSM4GOIHCFA