why am i getting error with this code?

testdata = readtable("testdata.xlsx");
predictions = char(trainedModel.predictFcn(testdata));

iscorrect=predictions==cell2mat(string((testdata.Character)));
iscorrect=iscorrect(:,2);
accuracy = sum(iscorrect)*100/30;

[Fifteen12]

What is the error you're getting? Without your test data it's difficult to replicate your issue.

[Walter Roberson]

Does the predictFcn of the trainedModel accept a table as input?

Why are you comparing only the second character of each predicted value against the table ?

Why are you dividing accuracy by the magic number 30 instead of by the number of rows in incorrect ?

[Kregozt]

I am trying to predict the accuracy between 2 models using this code ,am using iris dataset as my sample data. I have used code to get 80% trained data and 20% test data by splitting the dataset by using cvpartition function.

[Kregozt]

Everytime I run the code the error is arising at line 2 dont know why. I was referring classification learner app video by matlabhelper on Youtube and tried doing the same in matlab but the code was giving error

[Kregozt]

I was diving by 30 because from matlab documentation I learnt that division should be done by the number of rows in test data

[Walter Roberson]

accuracy = mean(iscorrect) * 100.

[Walter Roberson]

try

predictions = trainedModel.predictFcn(testdata);

and show us class(predictions) and size(predictions)

[Kregozt]

but i have done the same but only converting the same to character

[Walter Roberson]

It would make it easier for us to advise you if you were to call

predictions = trainedModel.predictFcn(testdata);
whos predictions

and show us the output of the whos call.

Answer 1

[Sulaymon Eshkabilov]

Without looking at your data, it is diffiuclt to address your question properly. Thus, pl., post your sample data.

[Kregozt]

I am using IRIS dataset downloaded from kraggle