Support Vector Machine – Regression (SVR)

Support Vector Machine – Regression (SVR)

Support Vector Machine (SVR) is a regression algorithm, so we can use SVR for working with continuous values instead of classification, which is SVM.
Support Vector Machine is one of the regression methods. Support Vector Machine maintains all the core features that describe the characteristics of the algorithm. The Support Vector Machine (SVM) for classification is mostly similar to the Support Vector Regression (SVR). However, only a little difference exists among both of these two methods.

Support Vector Machine - Regression (SVR)
Figure: Support Vector Machine – Regression (SVR)

Now, let’s see the most commonly used terms in this post tutorial.

Support vectors

Support vectors are the data points located nearly to the boundary. The distance of the points is minimum.


Mostly used to map data with lower dimension into data with a higher dimension.

Boundary line

Two lines which create a margin. The support vectors can be outside it or on the Boundary lines.

Hyper Plane

Hyper Plane is acting as the separation line between the data classes. Similarly, we can define it as the line that will help us to predict the target value. The target value is also called a continuous value.

Support Vector Machines belong to some specific class of algorithms. Support Vector Machines can be characterized by the followings;
1. The absence of local minima.
2. Kernels usage.
3. The sparseness of the solution and the acting on margin to control the capacity, or on a number of support vectors etc.

What is the main difference between a simple regression model and Support Vector Machines regression (SVR)?

In simple regression, we focus to minimize the error rate of data. However, in Support Vector Machine (SVR) our main focus is to fix the error within a certain threshold.