Support Vector Machine Regression Predict

I have this Airlines dataset. I used 11 yers for training and 1 for testing.

Once i've chosen and trained the model (using fitrsvm, predict...), suppose that i need to predict the next 3 months how can i do?

Probably it's a dumb question, but i can't figure out.

Thanks.

Answer 1

[Iuliu Ardelean]

Hey

You could try:

numTrainPoints = 100; % Let's say you want 100 train points
dtrain = d(1:numTrainPoints, 1:(A-af); % select first 100 points for training
model = fitrsvm(dtrain, dresponse); % fit model to train data
dtest = d(numTrainPoints:end, 1:(A-af); % select the rest for testing
predictedPassengers = predict(model, dtest); % test 

[TOMMASO BERTOCCHI]

Thanks, but doesn't work.

Here the messagge error...

Error using classreg.learning.internal.numPredictorsCheck (line 29)
X data must have 11 columns.

[Iuliu Ardelean]

Hey -- sorry about that -- my error -- I've updated the response -- please let me know if it still doesn't work

[TOMMASO BERTOCCHI]

again thanks but still doesn't work.

Sorry

Index in position 1 exceeds array bounds (must not exceed 12).

Error in SVM_Question (line 28)
dtrain = d(1:numTrainPoints, 1:(A-af)); % select first 100 points for training

[Iuliu Ardelean]

what value did you use for numTrainPoints?

Maybe try a smaller number, e.g. numTrainPoints = 10

[TOMMASO BERTOCCHI]

it works only with numTrainpoints = 12, but it gaves the same result of the benigging

numTrainPoints = 12;
dtrain = d(1:numTrainPoints, 1:(A-af)); % select first points for training
model = fitrsvm(dtrain, dresponse); % fit model to train data
dtest = d(1:numTrainPoints, 1:(A-af)); % select the rest for testing
predictedPassengers = predict(model, dtest); % test

because it's the same code dtest result the same of dtrain