Today I have started testing SVM regression models on the agriculture data that I’ve so far used for this year’s neural network publications. There are numerous implementations for SVM regression, some of which may be found at http://www.svms.org/software.html or http://www.support-vector-machines.org/SVM_soft.html.
After some research I figured out that SVMTorch suited my purposes quite well and hacked away a matlab script that does some k-fold cross validation (input data are split block-wise). In the end it generates mean absolute error and root mean squared error, just for starters.
I haven’t done any SVM fine-tuning, haven’t even set special parameters; yet, the model comes up with an MAE of 0.54 and RMSE of 0.73. For comparison: in one of my publications I ended up with an RMSE of 0.44 to 0.55, depending on the neural network’s dimensions. I’ll keep working with this SVM approach.