Support Vector Machines (SVMs) are methods used in Supervised learning, applied mainly to classification tasks.

SVMs try to find the most optimal separation line/plane between classes in a vector space.

Low dimensional data is transformed to a higher dimension to then find a Support Vector Classifier. The transformation that is used is determined by a Kernel Function.

Support Vectors are observations/data samples that are located in the Soft margin of that classification.


  • [b] Support Vector Classifiers (the baby SVMs)
  • [b] Kernel functions, e.g. polynomial and radial functions + Kernel tricks