Support Vector Machine (SVM) is a supervised learning algorithm for classification and regression. It finds the optimal hyperplane, a flat decision boundary, that separates data points into categories with the widest possible margin between classes. The data points closest to this boundary are called support vectors because they define its position. If the data cannot be separated by a flat boundary, SVM uses the kernel trick to map inputs into a higher-dimensional space where a linear separator exists, then projects the result back. SVM excels at high-dimensional data with relatively few training examples. In text classification, each word can be a dimension, resulting in thousands of features. SVM handles this without overfitting. In bioinformatics, it distinguishes disease-related gene expression patterns from healthy ones. Because the model depends only on support vectors, it stays compact and fast to evaluate. SVM remains competitive with deep neural networks on smaller datasets where deep models lack sufficient data to generalize. Industries from finance to autonomous vehicles rely on it for fraud detection, image recognition, and sensor fusion.
Back to Glossary