radzilu / arduino-svm Goto Github PK

Can we run Arduino-SVM in just arduino only without using processing?
I would like to classify audio obtained through microphone.

Something is wrong with the SVM being generated by Arduino-SVM. The results of svm_evaluate are the same regardless of the class of the data given, with the result that the prediction is always the same.

I'm struggling to understand and thus debug the code. In particular, I'm puzzled by the way that svm_predict is looking up coefficient data from the model file. The supportVectors and valuesForSupport arrays do not appear to be consistent with the model file – there appears to be an off by one error on line 183 of the Processing file. Either that or I am failing to understand how the resulting arrays are being used.

Hi. I'm working on a SVM project in Python and I'm trying to use an SVM in Aduino. I'm trying to follow this example, but I don't know how to create a model with training data.

I think it is a great project. so……