Study On The Algorithm Of Support Vector Regression And Its Applications In Data Analysis |
Posted on:2016-03-22 | Degree:Master | Type:Thesis |
Country:China | Candidate:Y Y Zou | Full Text:PDF |
GTID:2308330479989090 | Subject:Applied Mathematics |
Abstract/Summary: | PDF Full Text Request |
We study on the coeffcient matrix of the equation(3.3) of the multi-output support vector regression machine algorithm. And provided the estimation theorem of the upper bound of the condition number of the coe?cient matrix and detail of the theory justification. Meanwhile we verify the validity and e?ectiveness of the estimation theory by experimentation.The theory study of spherical function approximation is becoming a hot research area.It is based on this consideration that we researched the spherical approximation by support vector regression machine algorithm. Simultaneously, the common kernels on the sphere is defined. A constraint condition is added in the normal support vector regression machine,which makes it possible to fit the spherical function approximation. The theoretical result is right by the numerical test of examples. And the results show that achieved approximation e?ect by choosing proper kernel. |
Keywords/Search Tags: | SVRM, Multi-output, condition numbers, sphere SVR, approximation |
PDF Full Text Request |
Related items |