Introduction To Support Vector Machines and Applications

Introduction To Support Vector Machines and Applications, everything you should know before starting to understand machine learning in Bioinformatics.

SVMs: A New Generation of Learning Algorithms

Support vector machines are a kind of regulated machine calculation for realizing which is utilized for grouping and relapse errands. In spite of the fact that they are utilized for both characterization and relapse, they are basically utilized for arrangement challenges.

The first SVM calculation was designed by Vladimir N. Vapnik and Alexey Ya. Chervonenkis in 1963.

A help vector calculation is performed by plotting each procured estimation of information as a point on an n-dimensional space or chart. Here “n” speaks to the absolute number of a component of information that is available. The estimation of every datum is spoken to as a specific organization on the diagram.

After the conveyance of facilitating information, we can perform characterization by finding the line or hyper-plane that unmistakably partitions and separates between the two classes of information.


Support vector machines are an instrument that best effectively separates two classes. They are a part based calculation.

A portion alludes to a capacity that changes the information into a high dimensional space where the inquiry or issue can be settled.

A portion capacity can be either straight or non-direct. Portion techniques are a sort of class of calculations for design investigation.

The essential capacity of the part is to get information as info and change them into the necessary types of yield.

In insights, “piece” is the mapping capacity that ascertains and speaks to estimations of 2-dimensional information in a 3-dimensional space group.

A help vector machine utilizes a part stunt that changes the information to a higher measurement and afterward it attempts to locate an ideal hyperplane between the yields conceivable.

Piece’s strategy for the examination of information in help vector machine calculations utilizing a straight classifier to take care of non-direct issues is known as ‘bit stunt’.

Parts are utilized in measurements and math, however, it is most generally and furthermore most regularly utilized in help vector machines.


Pre 1980:

  • Almost all learning methods learned linear decision surfaces.
  • Linear learning methods have nice theoretical properties


  • Decision trees and NNs allowed efficient learning of nonlinear decision surfaces
  • Little theoretical basis and all suffer from local minima


  • Efficient learning algorithms for non-linear functions based on computational learning theory developed
  • Nice theoretical properties.

Key Ideas

Statistical Learning Theory

  • Systems can be mathematically described as a system that
    • Receives data (observations) as input and
    • Outputs a function that can be used to predict some features of future data.
  • Statistical learning theory models this as a function estimation problem
  • Generalization Performance (accuracy in labeling test data) is measured

The motivation for Support Vector Machines

  • The problem to be solved is one of the supervised binary classifications. That is, we wish to categorize new unseen objects into two separate groups based on their properties and a set of known examples, which are already categorized.
  • A good example of such a system is classifying a set of new documents into positive or negative sentiment groups, based on other documents that have already been classified as positive or negative.
  • Similarly, we could classify new emails into a spam or non-spam, based on a large corpus of documents that have already been marked as spam or non-spam by humans. SVMs are highly applicable to such situations.
  • A Support Vector Machine models the situation by creating a feature space, which is a finite-dimensional vector space, each dimension of which represents a “feature” of a particular object. In the context of spam or document classification, each “feature” is the prevalence or importance of a particular word.
  • The goal of the SVM is to train a model that assigns new unseen objects into a particular category.
  • It achieves this by creating a linear partition of the feature space into two categories.
  • Based on the features in the new unseen objects (e.g. documents/emails), it places an object “above” or “below” the separation plane, leading to a categorization (e.g. spam or non-spam). This makes it an example of a non-probabilistic linear classifier. It is non-probabilistic because the features in the new objects fully determine its location in feature space and there is no stochastic element involved.

Introduction To Support Vector Machines and Applications


  • Support vector machines(SVM) are supervised learning models with associated learning algorithms that analyze data used for classification and regression analysis.
  • It is a machine learning approach.
  • They analyze the large amount of data to identify patterns from them.
  • SVMs are based on the idea of finding a hyperplane that best divides a dataset into two classes, as shown in the image below.
Support vector machines(SVM)
  • Support Vectors are simply the coordinates of individual observation. Support Vector Machine is a frontier that best segregates the two classes (hyper-plane/ line).
  • Support vectors are the data points that lie closest to the decision surface (or hyperplane)
  • They are the data points most difficult to classify
  • They have a direct bearing on the optimum location of the decision surface
  • We can show that the optimal hyperplane stems from the function class with the lowest “capacity” (VC dimension).
  • Support vectors are the data points nearest to the hyperplane, the points of a data set that, if removed, would alter the position of the dividing hyperplane. Because of this, they can be considered the critical elements of a data set.

What is a hyperplane?

As a simple example, for a classification task with only two features, you can think of a hyperplane as a line that linearly separates and classifies a set of data.

Intuitively, the further from the hyperplane our data points lie, the more confident we are that they have been correctly classified. We, therefore, want our data points to be as far away from the hyperplane as possible, while still being on the correct side of it.

So when new testing data are added, whatever side of the hyperplane it lands will decide the class that we assign to it.

Introduction To Support Vector Machines and Applications

How do we find the right hyperplane?

How do we best segregate the two classes within the data?

The distance between the hyperplane and the nearest data point from either set is known as the margin. The goal is to choose a hyperplane with the greatest possible margin between the hyperplane and any point within the training set, giving a greater chance of new data being classified correctly. There will never be any data point inside the margin.

How do we find the right hyperplane?

But what happens when there is no clear hyperplane?

Data are rarely ever as clean as our simple example above. A dataset will often look more like the jumbled balls below which represent a linearly nonseparable dataset.

In order to classify a dataset like the one above it’s necessary to move away from a 2d view of the data to a 3d view. Explaining this is easiest with another simplified example.

Imagine that our two sets of colored balls above are sitting on a sheet and this sheet is lifted suddenly, launching the balls into the air. While the balls are up in the air, you use the sheet to separate them. This ‘lifting’ of the balls represents the mapping of data into a higher dimension. This is known as kernelling.


Because we are now in three dimensions, our hyperplane can no longer be a line. It must now be a plane as shown in the example above. The idea is that the data will continue to be mapped into higher and higher dimensions until a hyperplane can be formed to segregate it.

Support Vector Machine – Regression (SVR)

  • Support Vector Machine can also be used as a regression method, maintaining all the main features that characterize the algorithm (maximal margin).
  • The Support Vector Regression (SVR) uses the same principles as the SVM for classification, with only a few minor differences.
  • First of all, because the output is a real number it becomes very difficult to predict the information at hand, which has infinite possibilities.
  • In the case of regression, a margin of tolerance (epsilon) is set in approximation to the SVM which would have already requested from the problem. But besides this fact, there is also a more complicated reason, the algorithm is more complicated therefore to be taken into consideration.
  • However, the main idea is always the same: to minimize error, individualizing the hyperplane which maximizes the margin, keeping in mind that part of the error is tolerated.

Introduction To Support Vector Machines and Applications

Applications of SVM In Bioinformatics

SVMs depend as we have seen on supervised algorithms in learning. The objective of using SVM is to classify invisible data correctly. SVMs in various fields have a series of applications.

There are some common SVM applications:

Bioinformatics:- Protein remote homology detection is a common problem in the field of computational biology. SVM is the most effective way of solving this problem. SVM algorithms for the detection of remote protein homology have been extensively used for the last few years. These algorithms were widely used for biological sequence identification. For example, gene gradation, gene-based patients and numerous other biological issues.

Protein Fold and Remote Homology Detection:-

The determination of the three-dimensional (3D) design is key to understanding the function of biological macromolecules, e.g. proteins. A huge number of possible protein sequences are accumulating in large gene scale projects.

However, only a small fraction of known proteins can be provided with information on 3D structures. Thus, the sequence-structure gap continues to widen even if the experimental structure determination is improved.

In this way, structural information needs to be extracted from sequence databases. The direct prediction of the 3D structure of a protein from a sequence is not clear. There has however been substantial progress in assigning a folding class sequence.
This problem has been addressed in two general ways. One is the use of algorithms for threading. The second is a taxonomic approach presuming that the number of folds is limited and so structural predictions are focused on in a specific 3D fold classification.

The remote detection of protein homology is a major computer biology problem. One of the most effective methods for remote homology detection is supervised learning algorithms on SVMs. How the protein sequences modelled depends on the performance of the methods. The way the kernel function between them is computed.

Protein Secondary Structure Prediction:- Prognosis recently became ever more important for protein structure and function. The local conformation of the polypeptide chain, known as the secondary, is predicted as one step towards the full 3D structure. The secondary structure consists of hydrogen-bonded local plying regularity and traditionally consists of three classes: alpha-helix, beta-helixes and spindle.

A secondary structure was one of the classic problems in computational molecular biology due to the preferences and correlation in sequence involved, with a particularly successful machine learning approach.

Signal Peptide Cleavage Sites:- Procaryotic and eucaryotic cells are the target proteins of signal peptides. Signal recognition particle (SRP), which stops the translation, recognizes the signal peptide of the nascent protein on a free ribosome. SRP binds an endoplasmic (ER) membrane to an SRP receptor and inserts the signal peptide into the membrane.

Translation resumes and the protein is synthesized into the ER lumen via the membrane. Additional sequence determinants of the protein will then determine whether it will remain in the ER lumen, transmit or be released from the cell to one of the other membrane-bound compartments.

Signal peptides control the entrance into the secretory path of virtually all proteins. The N-term of the amino acid chain is the protein translocated across the membrane and is split off. There are three regions in the common structure of signal peptides, namely a positive n region, followed by a hydrophobe h-region and a neutral but polar c-region. The cleavage site is generally characterized by neutral small side-chain amino acids at positions -1 and-3 (relative to the cleavage site)

The large amount of data unprocessed available and the need to find more efficacious vehicles to produce proteins in recombinant systems evoked great interest in the forecast of signal peptids and their cleavage sites.

Stay tuned with us for more articles like this!!

Introduction To Support Vector Machines and Applications

Find more exciting articles in the Education section of this site.

Rajat Singh
Rajat Singh is the chief Author at Bioinformatics India, he has been writing for the past 3 years and has a special interest in SEO, Technology, Health, Life Sciences and gaming.

Get in Touch

Comments are closed.

Related Articles

Get in Touch


Subscribe to our newsletter

To be updated with all the latest news, offers and special announcements.

Latest Posts