The Support Vector Machine (SVM) is a powerful classification technique based on Kernels, such as the Kernel Ridge Regression (KRR) algorithm seen in the previous chapter. We often deal with sparse datasets or with data that is not good enough to make a good prediction or classification. In such cases, we may use a technique that creates new values from the original dataset to help in the accuracy of the algorithm; this new data is called synthetic. Due to their efficiency, using Kernels is one of the most common methods to make synthetic data. In this chapter, we will provide you with an easy way to get acceptable results using SVM. We will perform a dimensionality reduction of the dataset, and we will produce a model for classification.

The theoretical foundation of SVM lies in the work of Vladimir Vapnik and the theory of statistical learning developed in the 70s. SVMs are highly used in pattern recognition of Time Series, Bioinformatics...