Font Size: a A A

Support Vector Machine Classical Algorithm And Classification Research

Posted on:2012-10-18Degree:MasterType:Thesis
Country:ChinaCandidate:Z K LeiFull Text:PDF
GTID:2248330395965415Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Support Vector Machine (SVM) is a new powerful machine learning method which developed in the framework of Statistical Learning Theory (SLT).It has high generalization. At present, being the best theory for small samples learning, SLT and SVM are getting much emphasis from many fields,and have become a new research hot point in the fields of artificial intelligent and machine learning. In this paper, the contents are as follows:This paper, on the basic knowledge of support vector machine, highlights three classical algorithms:chunking, decomposing, sequential minimal optimization.It briefly describes the intrinsic relationships between the algorithms. Then by experiments, with two-dimensional eigenpostures datasets trains chunking and Sequential Minimal Optimization, using figures and tables show the experimental results, comparisons and analysises of performance of chunking and sequential minimal optimization.Then the paper discussed the problem of support vector classification, including the optimal separating hyperplane(linearly separable example), the generalized optimal separating hyperplane(linearly non-separable example), the generalization in high dimensional feature space (polynomial mapping example). Finally, IRIS data set show performance of support vector machine classification algorithm. People can directly feel the classification results.In the last chapter, I summarize the paper’s contents and propose some suggestions of the future work.
Keywords/Search Tags:support vector machine VC Dimension, Chunking DecomposingSequential Minimal Optimization, Support Vector Classification, IRIS Data Set
PDF Full Text Request
Related items