Support vector machine (SVM) is one of the simplest machine learning models. Understanding this is like learning basic to machine learning and progressing ahead. Not used too much but could be effective for solving “SVM” specific problems and while training SVM model, there will be choices to be made regarding data preprocessing, kernel selection and setting the parameter. I hope it is understood that how kernel selection depends on the problem statement to solve linear or non Linear solution.

Let us go to the basics of what vector is doing in SVM and what it is? — Vector is made up of one row or column of scalar values. eg, Scalar value is any number then Vector is [ 20 12 13] (row) and similarly we can have single column vector. Mutliple vector will create matrix. Just a refresher to have basic understanding before we deep dive into SVM.

Few basics about SVM: It is supervised machine learning algorithm. To clarify, supervised machine learning model means training data will be labelled while training the model. We can make a guess that why do we need labelled or pre classified training data. In simple, SVM has to find the support vectors as the name suggest to find the optimal solution. And, labelled data helps to achieve this.

#kernel-trick #svm #data-science #machine-learning-models #machine-learning