About decomposition approach for solving the classification problem

11th International Conference on "Mesh methods for boundary-value problems and applications"

Andrianova, A.A.^1

Department of System Analysis and Information Technologies, Kazan (Volga Region) Federal University, Kremlevskaya st. 18, Kazan, Russia^1

关键词: Binary classification problems;    Computational experiment;    Data set;    Decomposition approach;    Decomposition methods;    Linear classifiers;    Parallel version;    Support vector machine method;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/158/1/012006/pdf
DOI  :  10.1088/1757-899X/158/1/012006

来源: IOP

【 摘 要 】

This article describes the features of the application of an algorithm with using of decomposition methods for solving the binary classification problem of constructing a linear classifier based on Support Vector Machine method. Application of decomposition reduces the volume of calculations, in particular, due to the emerging possibilities to build parallel versions of the algorithm, which is a very important advantage for the solution of problems with big data. The analysis of the results of computational experiments conducted using the decomposition approach. The experiment use known data set for binary classification problem.