Support Vector Machine Accuracy Improvement with Classification

Abstract

Rapid increase in the information technology through digitalization, leads to fast enhancement in technical industry has expanded the need for effective data mining. Data mining is a combination of tools and mechanisms for taking out of beneficial data set from a gigantic amount of databases. Data mining has three main components: Classification, Sequence Analysis, and Association rule. Data mining provides various functions like; error identification, regression, clustering, classification, association rule, and summarization. This is a machine learning method used to estimate set membership for data instances. Machine learning is defined as the capability of a system to absorb from its past experience. Support vector machine is widely used to classify linear and nonlinear information and instigate the data based on traditional statistics theory. Support vector machine used to draw a decision surface that classifies all training vector into subclasses. There supervised machine learning model (SVM) used as a classifier as well as we tried to achieve better pattern recognition using classification technique of data mining. A comparison of different types of kernels performance using Support Vector Machine has also to be done. These results yield a good pattern prediction rate and low misclassification of data.