Introduction to Support Vector Machines

Support Vector Machines (SVM) is a machine learning technique based on statistical learning theory. It is used to classify different types of objects, and it has many application areas today. One of the most popular uses of Support Vector Machine is Optical Character Recognition (OCR).

SVM has some similarities to neural network classification. However, SVM uses simpler and therefore more comprehensible mathematical methods. This learning algorithm was developed by V. Vapnik for the classification of data. SVMs stand out because they are highly applicable to real problems and generalize well.

The Concept of Support Vector Machines

We use the following concepts for SVMs:

  • Low dimension data is mapped to a new feature space of higher dimension. Our machine then learns with this new high-dimensional data rather than the original data. This seems complicated at first, but you can classify the data optimally in the new room with much simpler, mostly linear methods.
  • Apply linear approximation functions to the new feature space.
  • The risk of incorrectly classifying data is minimized by using simple mathematical formulas to calculate the optimal separation.
  • The solution fits a dual optimization problem.



Did you like our introduction to support vector machines? You can find many more guides like like this one right here: Machine Learning Beginner.