Support Vector Machines (SVM) are a class of supervised learning algorithms often used for classification or sometimes regression tasks. Generally, you first need to build an SVM to either classify a set of samples into their respective classes (SVM for classification) or fit a set of data samples to an expected target (SVM for regression). In either of the above case, you need to use part of the entire data to train the built model and used the remaining data to test the trained model. As for the label, in the case of supervised learning (for classification problems), the labels should be decided by an expert.
Kindly see the following article for SVM application: "Adaptive context aware decision computing paradigm for intensive health care delivery in smart cities—A case analysis"
in addition to the comments above, you can also have a look at the ARDUOUS workshop series. They address the labelling of data needed for classification tasks. I think you can find a lot of useful materials there.