<< Click to Display Table of Contents >> Navigation:  Introduction > Architectures included in Distribution > Support Vector Machine (SVM)

The SVM architecture for classification in PAI is implemented in accordance with the Python library scikit-learn https://scikit-learn.org/stable/modules/generated/sklearn.svm.SVC.html

Support vector machines are supervised learning models used for classification and regression:

https://en.wikipedia.org/wiki/Support-vector_machine

SVM is particularly suited to image reduction strategies such as using the average voxel value in a set of VOIs (e.g. brain atlas VOIs for brain PET data - see Amyloid PET classification case study). Linear, RBF, Sigmoid and Poly kernels are available according to the scikit-learn library (see Learning Set Preparation). Our testing of the architecture in PAI has used up to 10 classes.

The output of a trained SVM classification model is a probability that an input to prediction belongs to one of the classes included in training.

An example of the SVM architecture in practice is provided in our Amyloid PET classification case study. Data to test the Amyloid PET SVM classification model is available in our Demo database.