Support Vector Machine - Simple Analysis on Classifying Data, Confusion Matrix and Kernel Selection

Abstract: Machine learning algorithms come under different categories like supervised, unsupervised and Reinforcement learning. In each case many algorithms are there that serves specific purposes like Regression, Classification, Clustering, and Dimensionality Reduction and so on. Regression type algorithms are mainly used for prediction, Classification is used for dividing or categorizing, Clustering is for grouping similar items. In Reinforcement learning it analyses and optimizes the behaviour of an agent based on the environment. We have to choose the correct algorithms available based on the output required. Because for a single problem we may have different solutions but the thing is we have to choose the one with high accuracy and it should fit for any updating in the datasets.

Keywords: Machine Learning, Classification, Kernel, Hyper Planes and Confusion Matrix